- Email: [email protected]
Fondaparinux and prevention of venous thromboembolism after orthopaedic surgery
CORRESPONDENCE
COMMENTARY
CORRESPONDENCE e-mail submissions to [email protected]
Fondaparinux and prevention of venous thromboembolism after orthopaedic surgery Sir—In their Commentary (Aug 16 p 504),1 Gordon Lowe and colleagues make several misleading points, which warrant critique. First, they call venogram-diagnosed postoperative asymptomatic deepvein thrombosis (DVT) a “clinically irrelevant surrogate”. Doing so is like calling a fortuitously discovered CT-diagnosed lung carcinoma or hepatoma irrelevant. Indeed, the surrogate of phlebographically diagnosed DVT remains the only one accepted by most regulatory agencies, including the US Food and Drug Administration (FDA), when registering a new thromboprophylactic agent. There is a good basis for this policy:2 asymptomatic, distal DVT is probably the only causal pathway that leads to proximal DVT and pulmonary embolism, and the reduction of adverse clinical events is due to the effect of antithrombotic drugs on the surrogates. In the case of phlebographically diagnosed DVT, fatal pulmonary embolism will not arise if all phlebographically diagnosed asymptomatic thrombi are prevented. Second, clinically relevant is not synonymous with symptomatic, as suggested by Lowe and colleagues. What is clinically more relevant, a painful calf DVT or an asymptomatic iliac vein thrombosis? Consider the difference between a painful peripheral pulmonary embolus and an asymptomatic central embolus. Of course, symptomatic might also mean clinically relevant—eg, a fatal pulmonary embolism. In clinical trials of efficacy, use of a valid surrogate endpoint is more pertinent than the biased criteria advocated by Lowe and co-workers, such as non-specific and insensitive symptoms like leg pain or swelling after total hip arthroplasty. HB has received departmental support from several pharmaceutical companies, including Sanofi-Synthelabo, for trials of antithrombotic therapies, and occasional honoraria and travel and accommodation support for lectures given.
Henri Bounameaux Division of Angiology and Haemostasis, University Hospital of Geneva, CH-1211 Geneva 14, Switzerland (e-mail: [email protected])
1
2
Lowe GDO, Sandercock PAG, Rosendaal FR. Prevention of venous thromboembolism after major orthopaedic surgery: is fondaparinux an advance? Lancet 2003; 362: 504–05. Fleming T, DeMets DL. Surrogate end points in clinical trials: are we being misled? Ann Intern Med 1996; 125: 605–13.
Sir—Gordon Lowe and colleagues1 have several concerns about the clinical introduction of fondaparinux, and I endorse their overall conclusions. Indeed, the impression one gets is that the series of publications and associated meta-analyses by investigators who were also involved in the design and undertaking of the original fondaparinux studies has been part of a well orchestrated plan aimed at swift introduction of this antithrombotic agent to the market. All involved, including investigators well respected in the specialty and medical journals have been, willingly or not, part of this aim. Given these concerns, the major question remains: how will such matters be dealt with now and in the future? Unfortunately, not all the recommendations made by Lowe and colleagues are helpful. One piece of advice is to acquire “much further clinical data and experience” with fondaparinux. Such advice applies to every new drug, but the main question for orthopaedic surgeons and pharmacists will be whether this agent should be introduced, and that decision can only be made on the basis of available evidence. If the authors believe that fondaparinux has not been sufficiently validated, they should not encourage its use. The second recommendation made concerns the use of clinical versus surrogate endpoints, and I agree that such endpoints should be analysed separately. Clinical endpoints are of importance; unfortunately the major burden in orthopaedic surgery originates from clinically silent venous thromboembolism (VTE). Since surrogate endpoints are associated with clinical endpoints in this category of patients, there are no strong arguments to reject the use of surrogate markers in this instance. That such endpoints could disguise clinically important data, however, should be remembered. All these arguments circumvent the
main difficulty—namely, should phase III trials remain exclusively companysponsored? The debate on oral contraceptives and the associated risk of venous thrombosis emphasised the differences that can arise in trials sponsored by industry and in those sponsored by other organisations.2 If company-sponsored phase III trials generate data that cannot be trusted, surely they should be stopped? In view of the huge investments that companies need to make to develop clinically useful drugs, perhaps future customers—ie, those involved in health care and patients—should take on some of the responsibility by participating in study design and funding. In the USA, the National Institutes of Health (NIH) sponsors some drug trials. Likewise, European health institutes should perhaps consider participating in phase III trials with new drugs by funding in part or fully the trial, in collaboration with those involved in industry and clinical research. An independent council should supervise the entire process of planning, design of the study, and its undertaking, which would protect those involved in the process from potential conflicts of interest. Obviously, such measures would pose a financial burden for the national health-care systems, but this cost could be defended on the basis that the introduction of new drugs poses an increasing burden on society already. If, due to flaws in trial design and analysis, the expenses for new drug prescription are not properly evidence-based, any attempt to obtain more reliable evidence is justified and should be invested in. HtC has received honoraria for consultation and presentations from different pharmaceutical companies.
Hugo ten Cate Department of Internal Medicine and Cardiovascular Research Institute Maastricht, University Hospital Maastricht, Maastricht, Netherlands (e-mail: [email protected]) 1
2
Lowe GDO, Sandercock PAG, Rosendaal FR. Prevention of venous thromboembolism after major orthopaedic surgery: is fondaparinux an advance? Lancet 2003; 362: 504–05. Vandenbroucke JP, Helmerhorst FM, Rosendaal FR. Competing interests and controversy about third generation oral contraceptives. BMJ 2000; 320: 381–82.
THE LANCET • Vol 362 • November 8, 2003 • www.thelancet.com
For personal use. Only reproduce with permission from The Lancet.
1581
COMMENTARY
CORRESPONDENCE e-mail submissions to [email protected]
Fondaparinux and prevention of venous thromboembolism after orthopaedic surgery Sir—In their Commentary (Aug 16 p 504),1 Gordon Lowe and colleagues make several misleading points, which warrant critique. First, they call venogram-diagnosed postoperative asymptomatic deepvein thrombosis (DVT) a “clinically irrelevant surrogate”. Doing so is like calling a fortuitously discovered CT-diagnosed lung carcinoma or hepatoma irrelevant. Indeed, the surrogate of phlebographically diagnosed DVT remains the only one accepted by most regulatory agencies, including the US Food and Drug Administration (FDA), when registering a new thromboprophylactic agent. There is a good basis for this policy:2 asymptomatic, distal DVT is probably the only causal pathway that leads to proximal DVT and pulmonary embolism, and the reduction of adverse clinical events is due to the effect of antithrombotic drugs on the surrogates. In the case of phlebographically diagnosed DVT, fatal pulmonary embolism will not arise if all phlebographically diagnosed asymptomatic thrombi are prevented. Second, clinically relevant is not synonymous with symptomatic, as suggested by Lowe and colleagues. What is clinically more relevant, a painful calf DVT or an asymptomatic iliac vein thrombosis? Consider the difference between a painful peripheral pulmonary embolus and an asymptomatic central embolus. Of course, symptomatic might also mean clinically relevant—eg, a fatal pulmonary embolism. In clinical trials of efficacy, use of a valid surrogate endpoint is more pertinent than the biased criteria advocated by Lowe and co-workers, such as non-specific and insensitive symptoms like leg pain or swelling after total hip arthroplasty. HB has received departmental support from several pharmaceutical companies, including Sanofi-Synthelabo, for trials of antithrombotic therapies, and occasional honoraria and travel and accommodation support for lectures given.
Henri Bounameaux Division of Angiology and Haemostasis, University Hospital of Geneva, CH-1211 Geneva 14, Switzerland (e-mail: [email protected])
1
2
Lowe GDO, Sandercock PAG, Rosendaal FR. Prevention of venous thromboembolism after major orthopaedic surgery: is fondaparinux an advance? Lancet 2003; 362: 504–05. Fleming T, DeMets DL. Surrogate end points in clinical trials: are we being misled? Ann Intern Med 1996; 125: 605–13.
Sir—Gordon Lowe and colleagues1 have several concerns about the clinical introduction of fondaparinux, and I endorse their overall conclusions. Indeed, the impression one gets is that the series of publications and associated meta-analyses by investigators who were also involved in the design and undertaking of the original fondaparinux studies has been part of a well orchestrated plan aimed at swift introduction of this antithrombotic agent to the market. All involved, including investigators well respected in the specialty and medical journals have been, willingly or not, part of this aim. Given these concerns, the major question remains: how will such matters be dealt with now and in the future? Unfortunately, not all the recommendations made by Lowe and colleagues are helpful. One piece of advice is to acquire “much further clinical data and experience” with fondaparinux. Such advice applies to every new drug, but the main question for orthopaedic surgeons and pharmacists will be whether this agent should be introduced, and that decision can only be made on the basis of available evidence. If the authors believe that fondaparinux has not been sufficiently validated, they should not encourage its use. The second recommendation made concerns the use of clinical versus surrogate endpoints, and I agree that such endpoints should be analysed separately. Clinical endpoints are of importance; unfortunately the major burden in orthopaedic surgery originates from clinically silent venous thromboembolism (VTE). Since surrogate endpoints are associated with clinical endpoints in this category of patients, there are no strong arguments to reject the use of surrogate markers in this instance. That such endpoints could disguise clinically important data, however, should be remembered. All these arguments circumvent the
main difficulty—namely, should phase III trials remain exclusively companysponsored? The debate on oral contraceptives and the associated risk of venous thrombosis emphasised the differences that can arise in trials sponsored by industry and in those sponsored by other organisations.2 If company-sponsored phase III trials generate data that cannot be trusted, surely they should be stopped? In view of the huge investments that companies need to make to develop clinically useful drugs, perhaps future customers—ie, those involved in health care and patients—should take on some of the responsibility by participating in study design and funding. In the USA, the National Institutes of Health (NIH) sponsors some drug trials. Likewise, European health institutes should perhaps consider participating in phase III trials with new drugs by funding in part or fully the trial, in collaboration with those involved in industry and clinical research. An independent council should supervise the entire process of planning, design of the study, and its undertaking, which would protect those involved in the process from potential conflicts of interest. Obviously, such measures would pose a financial burden for the national health-care systems, but this cost could be defended on the basis that the introduction of new drugs poses an increasing burden on society already. If, due to flaws in trial design and analysis, the expenses for new drug prescription are not properly evidence-based, any attempt to obtain more reliable evidence is justified and should be invested in. HtC has received honoraria for consultation and presentations from different pharmaceutical companies.
Hugo ten Cate Department of Internal Medicine and Cardiovascular Research Institute Maastricht, University Hospital Maastricht, Maastricht, Netherlands (e-mail: [email protected]) 1
2
Lowe GDO, Sandercock PAG, Rosendaal FR. Prevention of venous thromboembolism after major orthopaedic surgery: is fondaparinux an advance? Lancet 2003; 362: 504–05. Vandenbroucke JP, Helmerhorst FM, Rosendaal FR. Competing interests and controversy about third generation oral contraceptives. BMJ 2000; 320: 381–82.
THE LANCET • Vol 362 • November 8, 2003 • www.thelancet.com
For personal use. Only reproduce with permission from The Lancet.
1581