- Email: [email protected]
Tranexamic acid for trauma – Authors' reply
Correspondence
1
2 3
4
5
CRASH-2 trial collaborators. Effects of tranexamic acid on death, vascular occlusive events, and blood transfusion in trauma patients with significant haemorrhage (CRASH-2): a randomised, placebo-controlled trial. Lancet 2010; 376: 23–32. Levy JH. Antifibrinolytic therapy: new data and new concepts. Lancet 2010; 376: 3–4. Murkin JM, Falter F, Granton J, et al. High-dose tranexamic acid is associated with nonischemic clinical seizures in cardiac surgical patients. Anesth Analg 2010; 110: 350–53. Casati V, Della Valle P, Benussi S, et al. Effects of tranexamic acid on postoperative bleeding and related hematochemical variables in coronary surgery: comparison between onpump and off-pump techniques. J Thorac Cardiovasc Surg 2004; 128: 83–91. Casati V, Romano A, D’Angelo A. Safety of aprotinin (again). Lancet 2008; 372: 27.
We were very interested in the CRASH-2 trial.1 It showed not only the efficacy of tranexamic acid, but also its excellent safety in a large cohort of patients. Although the inclusion criteria for this trial were very relevant to routine clinical practice, hypotension in a trauma patient can be related to other causes (pneumothorax, spinal cord injury, etc) and not necessarily to documented bleeding.2 The CRASH-2 trial included younger patients than in other series and therefore underestimates pre-existing occlusive lesions with their corollary of platelet antiaggregant or anticoagulant treatments that predispose to bleeding. A high frequency of coagulopathy is established on emergency-room admission and is affected by prehospital management and the magnitude of injury (reflected by the injury severity score [ISS]). ISS is not presented in the results of the CRASH-2 trial. Organisation of the health-care system therefore has a major role.3 The CRASH-2 trial does not take into account prehospital organisation, raising the question of whether these results can be extrapolated to all healthcare systems. Since randomisation did not take into account patients in whom tranexamic acid would have been immediately indicated, this trial 1050
excluded the most critically ill patients. Many European and North American studies on the therapeutic approach to haemorrhage have assessed the effect on mortality of the transfusion ratio of fresh frozen plasma to packed red blood cells.4 The CRASH-2 trial did not address this issue. Although this trial done on an intention-to-treat basis has a substantial statistical power, the absence of analysis of the transfusion ratio could represent a bias and a limitation to the use of tranexamic acid for the most seriously ill patients. Finally, the prevalence of thrombotic complications in the CRASH-2 trial remains relatively low compared with other studies.5 Owing to its power and rigorous methods, this trial nevertheless remains a major trial in the therapeutic approach to haemorrhage in trauma patients, showing that this inexpensive and easy-to-use molecule could be particularly useful in emergency situations without any major adverse effects. We declare that we have no conflicts of interest.
*Jean-Michel Yeguiayan, Nadia Rosencher, Marc Freysz [email protected] Université de Bourgogne, CHU de Dijon, Service d’Anesthésie Réanimation - SAMU 21, 21033 Dijon, France (JMY, MF); and Université Paris Descartes, Hôpital Cochin, APHP, Paris, France (NR) 1
2
3
4
5
CRASH-2 trial collaborators. Effects of tranexamic acid on death, vascular occlusive events, and blood transfusion in trauma patients with significant haemorrhage (CRASH-2): a randomised, placebo-controlled trial. Lancet 2010; 376: 23–32. Carli PA, de la Coussaye JE, Riou B. Immediate versus delayed fluid resuscitation in patients with trauma. N Engl J Med 1995; 332: 682. Nathens AB, Brunet FP, Maier RV. Development of trauma systems and effect on outcomes after injury. Lancet 2004; 363: 1794–801. Rossaint R, Bouillon B, Cerny V, et al. Management of bleeding following major trauma: an updated European guideline. Crit Care 2010; 14: R52. Geerts WH, Jay RM, Code KI, et al. A comparison of low-dose heparin with lowmolecular-weight heparin as prophylaxis against venous thromboembolism after major trauma. N Engl J Med 1996; 335: 701–07.
Authors’ reply Biswadev Mitra and colleagues point out the low cost of tranexamic acid, which we showed can reduce mortality in bleeding trauma patients. This point is important since most trauma deaths are in low-income and middleincome countries. A cost-effectiveness analysis is in preparation. Of Mitra’s list of treatment options for traumatic haemorrhage, tranexamic acid is the only option for which there is reliable evidence of effectiveness from randomised controlled trials. We do not agree that the mortality rate of 15% implies that the CRASH-2 trial was done in underdeveloped trauma systems. The trial recruited patients from hospitals in five continents and this helps the results to be generalised. As regards coagulopathy, we made it clear in our paper that the inclusion criteria were clinical and did not include laboratory tests to assess coagulopathy. We agree with Valter Casati and colleagues that tranexamic acid doses much higher than those used in the CRASH-2 trial should not be routinely used in trauma patients outside the context of a clinical trial. In the CRASH-2 trial there were no reports of serious unexpected adverse events involving seizures. We agree with Jean-Michel Yeguiayan and colleagues that hypotension in trauma patients can be due to causes other than bleeding. We also agree that vascular occlusive lesions are more common in older patients. Although we did not find any increase in non-fatal vascular occlusive events with tranexamic acid in our trial, we acknowledge that we could not exclude the possibility of some increase in risk. We also acknowledged that it is possible that we might have under-reported the frequency of vascular occlusive events but point out that even if we had, our estimates of the relative risk of such events should be unbiased. We do not understand how the absence of an analysis of transfusion ratio could have introduced bias. We declare that we have no conflicts of interest.
www.thelancet.com Vol 376 September 25, 2010
Correspondence
[email protected] London School of Hygiene and Tropical Medicine, London WC1E 7HT, UK
Effects of fibrates on cardiovascular outcomes Min Jun and colleagues’ meta-analysis on the effects of fibrates on cardiovascular outcomes (May 29, p 1875)1 deserves comment. Jun and colleagues indicate that “For high-risk populations, a proportional risk reduction of 10–15% would translate into a worthwhile absolute risk reduction and a plausible number needed to treat.” But: (1) the plausible number for a concrete situation was not calculated; (2) there are no effects on all-cause and on cardiovascular mortality; (3) the problem of compliance and competition with other more important treatments in these polypharmacy patients was avoided; and (4) a more effective alternative exists—ie, statins. I declare that I have no conflicts of interest.
clearance). A potential explanation could be increased creatinine production. Creatininuria rises in parallel to serum creatinine.2 The two largest trials— ACCORD3 and FIELD4—of the three studies included in the “progression of albuminuria” meta-analysis used the ACR as a measurement of albuminuria.1 But if creatininuria is increasing owing to a non-renal effect of fibrates, the ACR will automatically decrease irrespective of any direct effect of albuminuria. If the rise in serum creatinine is reversible after stopping fenofibrate therapy,4 it would be interesting to know whether creatininuria and albuminuria (in μg/ min and not in mg/g creatinine) were modified concomitantly. We need to understand better the reason why serum creatinine is increasing with fibrates. We also need trials to study the effect of fibrates (and other lipid-lowering agents) either on true measurements of glomerular filtration rate or on hard endpoints such as the need for dialysis. We declare that we have no conflicts of interest.
*Pierre Delanaye, André J Scheen [email protected]
Public Health, 27 rue Voiture, 80000 Amiens, France
Department of Nephrology, Dialysis, Transplantation (PD), and Division of Diabetes, Nutrition and Metabolic Disorders and Clinical Pharmacology Unit (AJS), University of Liège, CHU Sart Tilman, 4000 Liège, Belgium
1
1
Alain Braillon [email protected] Jun M, Foote C, Lv J, et al. Effects of fibrates on cardiovascular outcomes: a systematic review and meta-analysis. Lancet 2010; 375: 1875–84.
2
In their meta-analysis,1 Min Jun and colleagues show that fibrate therapy has a beneficial effect on progression of albuminuria (14% reduction vs placebo; p=0·028). However, this potential positive renal effect is based on a surrogate marker (the albuminuria to creatininuria ratio [ACR]), and, we believe, could be an artifact. The positive renal effect above is seen even if serum creatinine increases in patients treated with fibrates. This creatinine increase is linked neither to a reduction in glomerular filtration rate (no modification of inulin clearance) nor to an effect on tubular creatinine secretion (no modification of creatinine www.thelancet.com Vol 376 September 25, 2010
3
4
Jun M, Foote C, Lv J, et al. Effects of fibrates on cardiovascular outcomes: a systematic review and meta-analysis. Lancet 2010; 375: 1875–84. Hottelart C, el Esper N, Rose F, Achard JM, Fournier A. Fenofibrate increases creatininemia by increasing metabolic production of creatinine. Nephron 2002; 92: 536–41. Ginsberg HN, Elam MB, Lovato LC, et al. Effects of combination lipid therapy in type 2 diabetes mellitus. N Engl J Med 2010; 362: 1563–74. Keech A, Simes RJ, Barter P, et al. Effects of longterm fenofibrate therapy on cardiovascular events in 9795 people with type 2 diabetes mellitus (the FIELD study): randomised controlled trial. Lancet 2005; 366: 1849–61.
Authors’ reply Alain Braillon raises the important issue of absolute treatment effects for people treated with fibrates. Since the studies included in our analysis included a range of different participant populations, substantial variation in the event rates was seen
(1·2–3·3% per year), although the relative risk reductions produced were remarkably consistent. On the basis of these reported event rates, the absolute risk reduction for major cardiovascular events over 5 years would range from 0·60% to 1·65% and the number needed to treat to prevent one cardiovascular event (NNT) would range from 60·6 to 166·7. As described in our paper, several trials have suggested that the relative risk reduction might be substantially greater in participants with high triglyceride concentrations and low HDL cholesterol concentrations, in which case the absolute effect size would be substantially greater and the NNT correspondingly smaller in this subpopulation. We agree that statin-based lipid modulation should remain a priority given its clear efficacy, but also believe that fibrate therapy could have a role in high-risk individuals. Braillon is correct that the effects of fibrates on top of routine statin therapy have been studied only by ACCORD,1 in which the effects of treatment were consistent with those seen in other trials. We concur with Pierre Delanaye and André Scheen that a better understanding of the renal effects of these agents is needed. The increase in creatinine seen with fibrates has been well described, but recent data from FIELD2 showed a lower creatinine concentration in fibrate-treated patients than in controls after active treatment was stopped. The apparent protective effect in retinopathy seen in our study would also support a microvascular benefit for fibrates. The analyses suggested by Delanaye and Scheen would indeed be of interest but are not possible on the basis of present published data.
Science Photo Library
*Haleema Shakur, Ian Roberts, Pablo Perel, on behalf of the CRASH-2 Trial Collaborators
VP reports receiving grants for clinical trials from Baxter, Johnson and Johnson, Novartis, Roche, and Servier; and lecture fees from Abbott, AstraZeneca, Roche, and Servier. BN reports receiving consulting fees from Pfizer, Roche, and Takeda; grant support from Johnson and Johnson, Merck Schering Plough, Servier, and United Healthcare Group; lecture fees and travel reimbursements from Amgen,
1051
1
2 3
4
5
CRASH-2 trial collaborators. Effects of tranexamic acid on death, vascular occlusive events, and blood transfusion in trauma patients with significant haemorrhage (CRASH-2): a randomised, placebo-controlled trial. Lancet 2010; 376: 23–32. Levy JH. Antifibrinolytic therapy: new data and new concepts. Lancet 2010; 376: 3–4. Murkin JM, Falter F, Granton J, et al. High-dose tranexamic acid is associated with nonischemic clinical seizures in cardiac surgical patients. Anesth Analg 2010; 110: 350–53. Casati V, Della Valle P, Benussi S, et al. Effects of tranexamic acid on postoperative bleeding and related hematochemical variables in coronary surgery: comparison between onpump and off-pump techniques. J Thorac Cardiovasc Surg 2004; 128: 83–91. Casati V, Romano A, D’Angelo A. Safety of aprotinin (again). Lancet 2008; 372: 27.
We were very interested in the CRASH-2 trial.1 It showed not only the efficacy of tranexamic acid, but also its excellent safety in a large cohort of patients. Although the inclusion criteria for this trial were very relevant to routine clinical practice, hypotension in a trauma patient can be related to other causes (pneumothorax, spinal cord injury, etc) and not necessarily to documented bleeding.2 The CRASH-2 trial included younger patients than in other series and therefore underestimates pre-existing occlusive lesions with their corollary of platelet antiaggregant or anticoagulant treatments that predispose to bleeding. A high frequency of coagulopathy is established on emergency-room admission and is affected by prehospital management and the magnitude of injury (reflected by the injury severity score [ISS]). ISS is not presented in the results of the CRASH-2 trial. Organisation of the health-care system therefore has a major role.3 The CRASH-2 trial does not take into account prehospital organisation, raising the question of whether these results can be extrapolated to all healthcare systems. Since randomisation did not take into account patients in whom tranexamic acid would have been immediately indicated, this trial 1050
excluded the most critically ill patients. Many European and North American studies on the therapeutic approach to haemorrhage have assessed the effect on mortality of the transfusion ratio of fresh frozen plasma to packed red blood cells.4 The CRASH-2 trial did not address this issue. Although this trial done on an intention-to-treat basis has a substantial statistical power, the absence of analysis of the transfusion ratio could represent a bias and a limitation to the use of tranexamic acid for the most seriously ill patients. Finally, the prevalence of thrombotic complications in the CRASH-2 trial remains relatively low compared with other studies.5 Owing to its power and rigorous methods, this trial nevertheless remains a major trial in the therapeutic approach to haemorrhage in trauma patients, showing that this inexpensive and easy-to-use molecule could be particularly useful in emergency situations without any major adverse effects. We declare that we have no conflicts of interest.
*Jean-Michel Yeguiayan, Nadia Rosencher, Marc Freysz [email protected] Université de Bourgogne, CHU de Dijon, Service d’Anesthésie Réanimation - SAMU 21, 21033 Dijon, France (JMY, MF); and Université Paris Descartes, Hôpital Cochin, APHP, Paris, France (NR) 1
2
3
4
5
CRASH-2 trial collaborators. Effects of tranexamic acid on death, vascular occlusive events, and blood transfusion in trauma patients with significant haemorrhage (CRASH-2): a randomised, placebo-controlled trial. Lancet 2010; 376: 23–32. Carli PA, de la Coussaye JE, Riou B. Immediate versus delayed fluid resuscitation in patients with trauma. N Engl J Med 1995; 332: 682. Nathens AB, Brunet FP, Maier RV. Development of trauma systems and effect on outcomes after injury. Lancet 2004; 363: 1794–801. Rossaint R, Bouillon B, Cerny V, et al. Management of bleeding following major trauma: an updated European guideline. Crit Care 2010; 14: R52. Geerts WH, Jay RM, Code KI, et al. A comparison of low-dose heparin with lowmolecular-weight heparin as prophylaxis against venous thromboembolism after major trauma. N Engl J Med 1996; 335: 701–07.
Authors’ reply Biswadev Mitra and colleagues point out the low cost of tranexamic acid, which we showed can reduce mortality in bleeding trauma patients. This point is important since most trauma deaths are in low-income and middleincome countries. A cost-effectiveness analysis is in preparation. Of Mitra’s list of treatment options for traumatic haemorrhage, tranexamic acid is the only option for which there is reliable evidence of effectiveness from randomised controlled trials. We do not agree that the mortality rate of 15% implies that the CRASH-2 trial was done in underdeveloped trauma systems. The trial recruited patients from hospitals in five continents and this helps the results to be generalised. As regards coagulopathy, we made it clear in our paper that the inclusion criteria were clinical and did not include laboratory tests to assess coagulopathy. We agree with Valter Casati and colleagues that tranexamic acid doses much higher than those used in the CRASH-2 trial should not be routinely used in trauma patients outside the context of a clinical trial. In the CRASH-2 trial there were no reports of serious unexpected adverse events involving seizures. We agree with Jean-Michel Yeguiayan and colleagues that hypotension in trauma patients can be due to causes other than bleeding. We also agree that vascular occlusive lesions are more common in older patients. Although we did not find any increase in non-fatal vascular occlusive events with tranexamic acid in our trial, we acknowledge that we could not exclude the possibility of some increase in risk. We also acknowledged that it is possible that we might have under-reported the frequency of vascular occlusive events but point out that even if we had, our estimates of the relative risk of such events should be unbiased. We do not understand how the absence of an analysis of transfusion ratio could have introduced bias. We declare that we have no conflicts of interest.
www.thelancet.com Vol 376 September 25, 2010
Correspondence
[email protected] London School of Hygiene and Tropical Medicine, London WC1E 7HT, UK
Effects of fibrates on cardiovascular outcomes Min Jun and colleagues’ meta-analysis on the effects of fibrates on cardiovascular outcomes (May 29, p 1875)1 deserves comment. Jun and colleagues indicate that “For high-risk populations, a proportional risk reduction of 10–15% would translate into a worthwhile absolute risk reduction and a plausible number needed to treat.” But: (1) the plausible number for a concrete situation was not calculated; (2) there are no effects on all-cause and on cardiovascular mortality; (3) the problem of compliance and competition with other more important treatments in these polypharmacy patients was avoided; and (4) a more effective alternative exists—ie, statins. I declare that I have no conflicts of interest.
clearance). A potential explanation could be increased creatinine production. Creatininuria rises in parallel to serum creatinine.2 The two largest trials— ACCORD3 and FIELD4—of the three studies included in the “progression of albuminuria” meta-analysis used the ACR as a measurement of albuminuria.1 But if creatininuria is increasing owing to a non-renal effect of fibrates, the ACR will automatically decrease irrespective of any direct effect of albuminuria. If the rise in serum creatinine is reversible after stopping fenofibrate therapy,4 it would be interesting to know whether creatininuria and albuminuria (in μg/ min and not in mg/g creatinine) were modified concomitantly. We need to understand better the reason why serum creatinine is increasing with fibrates. We also need trials to study the effect of fibrates (and other lipid-lowering agents) either on true measurements of glomerular filtration rate or on hard endpoints such as the need for dialysis. We declare that we have no conflicts of interest.
*Pierre Delanaye, André J Scheen [email protected]
Public Health, 27 rue Voiture, 80000 Amiens, France
Department of Nephrology, Dialysis, Transplantation (PD), and Division of Diabetes, Nutrition and Metabolic Disorders and Clinical Pharmacology Unit (AJS), University of Liège, CHU Sart Tilman, 4000 Liège, Belgium
1
1
Alain Braillon [email protected] Jun M, Foote C, Lv J, et al. Effects of fibrates on cardiovascular outcomes: a systematic review and meta-analysis. Lancet 2010; 375: 1875–84.
2
In their meta-analysis,1 Min Jun and colleagues show that fibrate therapy has a beneficial effect on progression of albuminuria (14% reduction vs placebo; p=0·028). However, this potential positive renal effect is based on a surrogate marker (the albuminuria to creatininuria ratio [ACR]), and, we believe, could be an artifact. The positive renal effect above is seen even if serum creatinine increases in patients treated with fibrates. This creatinine increase is linked neither to a reduction in glomerular filtration rate (no modification of inulin clearance) nor to an effect on tubular creatinine secretion (no modification of creatinine www.thelancet.com Vol 376 September 25, 2010
3
4
Jun M, Foote C, Lv J, et al. Effects of fibrates on cardiovascular outcomes: a systematic review and meta-analysis. Lancet 2010; 375: 1875–84. Hottelart C, el Esper N, Rose F, Achard JM, Fournier A. Fenofibrate increases creatininemia by increasing metabolic production of creatinine. Nephron 2002; 92: 536–41. Ginsberg HN, Elam MB, Lovato LC, et al. Effects of combination lipid therapy in type 2 diabetes mellitus. N Engl J Med 2010; 362: 1563–74. Keech A, Simes RJ, Barter P, et al. Effects of longterm fenofibrate therapy on cardiovascular events in 9795 people with type 2 diabetes mellitus (the FIELD study): randomised controlled trial. Lancet 2005; 366: 1849–61.
Authors’ reply Alain Braillon raises the important issue of absolute treatment effects for people treated with fibrates. Since the studies included in our analysis included a range of different participant populations, substantial variation in the event rates was seen
(1·2–3·3% per year), although the relative risk reductions produced were remarkably consistent. On the basis of these reported event rates, the absolute risk reduction for major cardiovascular events over 5 years would range from 0·60% to 1·65% and the number needed to treat to prevent one cardiovascular event (NNT) would range from 60·6 to 166·7. As described in our paper, several trials have suggested that the relative risk reduction might be substantially greater in participants with high triglyceride concentrations and low HDL cholesterol concentrations, in which case the absolute effect size would be substantially greater and the NNT correspondingly smaller in this subpopulation. We agree that statin-based lipid modulation should remain a priority given its clear efficacy, but also believe that fibrate therapy could have a role in high-risk individuals. Braillon is correct that the effects of fibrates on top of routine statin therapy have been studied only by ACCORD,1 in which the effects of treatment were consistent with those seen in other trials. We concur with Pierre Delanaye and André Scheen that a better understanding of the renal effects of these agents is needed. The increase in creatinine seen with fibrates has been well described, but recent data from FIELD2 showed a lower creatinine concentration in fibrate-treated patients than in controls after active treatment was stopped. The apparent protective effect in retinopathy seen in our study would also support a microvascular benefit for fibrates. The analyses suggested by Delanaye and Scheen would indeed be of interest but are not possible on the basis of present published data.
Science Photo Library
*Haleema Shakur, Ian Roberts, Pablo Perel, on behalf of the CRASH-2 Trial Collaborators
VP reports receiving grants for clinical trials from Baxter, Johnson and Johnson, Novartis, Roche, and Servier; and lecture fees from Abbott, AstraZeneca, Roche, and Servier. BN reports receiving consulting fees from Pfizer, Roche, and Takeda; grant support from Johnson and Johnson, Merck Schering Plough, Servier, and United Healthcare Group; lecture fees and travel reimbursements from Amgen,
1051