The n e w e ng l a n d j o u r na l of m e dic i n e
edi t or i a l s
New Anticoagulants — The Path from Discovery
to Clinical Practice
Jens Lohrmann, M.D., and Richard C. Becker, M.D.
For more than half a century, heparin and vita- symptomatic and asymptomatic venous thrombo-
min K antagonists have defined anticoagulant embolism and major venous thromboembolism,
therapy for the short-term and long-term man- defined as a composite of proximal deep-vein
agement, respectively, of thrombotic disorders of thrombosis, nonfatal pulmonary embolus, or death
the venous system. The history of their develop- related to venous thromboembolism. The frequen-
ment is instructive. In 1922, at the annual meet- cy of major bleeding and other safety outcomes
ing of the American Physiological Society, William — including on-treatment bleeding, hemorrhag-
H. Howell of Johns Hopkins Medical School pre- ic wound complications, and hepatic enzyme el-
sented an extraction protocol for isolating heparin evations — was low and did not differ between
preparations. Dicoumarol, a bacterial antagonist the study groups.
of vitamin K in spoiled sweet clover, was recog- The favorable observations regarding the use
nized as the agent responsible for a fatal hemor- of rivaroxaban raise questions that reach far be-
rhagic disease in livestock by Karl Link and Wil- yond individual drugs to the overall process of
helm Schoeffel of the University of Wisconsin. In drug development. What are the guiding princi-
both cases, decades would pass before heparin and ples for developing safe, effective, and widely ap-
warfarin entered clinical practice, and even today plicable anticoagulants? What properties, charac-
the complex pharmacokinetics, pharmacodynam- teristics, and outcomes that are discerned through
ics, and optimal use of these anticoagulants are clinical trials will help clinicians to distinguish
causes of uncertainty in the medical community.1 factor Xa inhibitors from one another should the
Two articles in this issue of the Journal describe future offer multiple options?
the results of thromboprophylaxis with an orally The environment in which the thrombus oc-
active, highly selective, direct inhibitor of factor curs is important. Biologic and physiological dis-
Xa, rivaroxaban, given at a fixed dose of 10 mg tinctions among the arterial, venous, and micro-
daily, as compared with the results with enoxapa- vascular systems are well known, as are intrinsic
rin, a subcutaneously administered, indirect, non- patient-specific differences in thromboresistance.
selective factor Xa inhibitor, in patients undergo- Indeed, the regulation of platelet–vessel wall in-
ing major orthopedic surgery. Collectively, more teractions, coagulation proteases, and fibrinolytic
than 7000 patients were included in the Regula- systems occurs on the endothelial surface, sug-
tion of Coagulation in Orthopedic Surgery to Pre- gesting that overall hemostatic regulation and the
vent Deep Venous Thrombosis and Pulmonary location, extent, and stability of thrombus forma-
Embolism (RECORD1) (ClinicalTrials.gov number, tion are conditionally based on site-specific dif-
NCT00329628)2 and RECORD3 (ClinicalTrials.gov ferences in endothelial-cell structure, function,
number, NCT00361894)3 studies, with similar re- and molecular responses to biologic and rheologi-
sults for total hip and total knee arthroplasty, re- cal conditions. The disease state (e.g., cancer) and
spectively. As compared with enoxaparin, rivaroxa- clinical circumstance (e.g., surgery) exert profound
ban was associated with significant reductions in effects on the site of predisposition to thrombo-
n engl j med 358;26 www.nejm.org june 26, 2008 2827
The n e w e ng l a n d j o u r na l of m e dic i n e
sis and the most effective target for intervention. phase 2 program that fosters drug-dose selection
After major orthopedic surgery, there is soft-tissue for phase 3 trials, an equally robust phase 3 pro-
injury, bone injury with intravasation of bony frag- gram, and a commitment to postmarketing safety
ments, and stasis of venous blood flow associat- surveillance are absolute prerequisites for success.
ed with surgical positioning of the limb. Surges of In turn, investigators must show great rigor and
tissue-factor release occur, and endothelial injury consistency in designing clinical studies and de-
that is caused by the intravasation of bone-cement fining end points.
constituents, such as methyl methacrylate mono- Even the safest and most effective anticoagu-
mer, contribute to the highly thrombotic envi- lant may not be used or may be used improperly
ronment. in clinical practice. For example, thromboprophy-
Molecular signatures and biomarkers that are laxis after orthopedic surgery, despite its efficacy
derived from genomics, pharmacogenomics, tran- and wide recommendation, is offered to less than
scriptomics, proteomics, and metabolomics are 60% of patients.9 Electronic alerts for anticoag-
likely to constitute the frontier for an improved ulation do increase the rate of prophylaxis,10 but
understanding of pharmacotherapy as it applies they are not 100% effective,11 which suggests that
to individual patients and disease.4,5 These tools an integrated approach, up to the level of “order
have the potential to inform patient care and make entry” and strict performance metrics, will be re-
a strong case for conducting parallel mechanis- quired to achieve greater success and optimal pa-
tic studies in phase 2 and 3 clinical trials. Un- tient care.
derstanding the disease, the drug,6 and the pa- Only through continued investigation and un-
tient offers a basis for the all-important selection derstanding of thrombosis and hemostasis can
of an appropriate dose and directs the physician progress in anticoagulant therapy be realized.
toward monitoring disease response rather than Myths, such as “an effective anticoagulant will,
drug response — the traditional end point of ther- by definition, carry with it a risk of bleeding” or
apy with heparin and drugs of similar complexity “it is easier to give a transfusion than to treat a
and off-target potential. blood clot” or “monitoring of an anticoagulant
Highly desirable properties of an anticoagu- will prevent its uptake in clinical practice,” are
lant are selectivity for the chosen hemostatic tar- obstacles that must be pushed aside to make way
get and the thrombotic disorder, rapidity of on- for change. The path to safer and more effective
set, capacity for a patient-specific regimen, and anticoagulants is paved by scientific knowledge,
rapid reversibility in the case of bleeding or a need discovery, due diligence on the part of sponsors
for a procedure requiring complete hemostasis. working collaboratively with experienced clini-
An orally active drug with these characteristics cians, and evidence-based translation to patient
offers flexibility for both acute and long-term ad- care and widespread clinical practice.
ministration. Gone are the days of nonselective
Dr. Becker reports receiving research grants from AstraZeneca,
anticoagulants with unfavorable pharmacokinet- Bayer, and Momenta. No other potential conflict of interest rel-
ics and pharmacodynamics, archaic and highly evant to this article was reported.
vulnerable manufacturing processes,7,8 and pre-
From the Division of Cardiology, University of Freiburg, Freiburg,
dictably unpredictable off-target effects. Germany (J.L.); and the Cardiovascular Thombosis Center, Duke
The transition from research on new drugs to University Medical Center, Durham, NC (R.C.B.).
the care of patients requires an enriched pipeline
1. Wardrop D, Keeling D. The story of the discovery of heparin
of drug discoveries, an infrastructure to foster the and warfarin. Br J Haematol 2008;141:757-63.
development of such drugs through early-phase 2. Eriksson BI, Borris LC, Friedman RJ, et al. Rivaroxaban ver-
testing, and an experienced and cohesive network sus enoxaparin for thromboprophylaxis after hip arthroplasty.
N Engl J Med 2008;358:2765-75.
of clinician investigators who are committed to 3. Lassen MR, Ageno W, Borris LC, et al. Rivaroxaban versus
conducting high-quality clinical research in a enoxaparin for thromboprophylaxis after total knee arthroplas-
timely fashion. The National Institutes of Health ty. N Engl J Med 2008;358:2776-86.
4. Becker RC. A rationale for conducting parallel mechanistic
Roadmap for Research (http://nihroadmap.nih. studies in clinical trials of pharmacotherapy. J Thromb Throm-
gov) is an example of a national biomedical re- bolysis (in press).
search enterprise that serves as a foundation for 5. Potti A, Bild A, Dressman HK, Lewis DA, Nevins JR, Ortel
TL. Gene-expression patterns predict phenotypes of immune-
improving clinical trials. In the case of trials spon- mediated thrombosis. Blood 2006;107:1391-6.
sored by the pharmaceutical industry, a robust 6. Mueck W, Eriksson BI, Bauer KA, et al. Population pharma-
2828 n engl j med 358;26 www.nejm.org june 26, 2008
cokinetics and pharmacodynamics of rivaroxaban — an oral, embolism risk and prophylaxis in the acute hospital care setting
direct factor Xa inhibitor — in patients undergoing major ortho- (ENDORSE study): a multinational cross-sectional study. Lancet
paedic surgery. Clin Pharmacokinet 2008;47:203-16. 2008;371:387-94.
7. Guerrini M, Beccati D, Shriver Z, et al. Oversulfated chon- 10. Kucher N, Koo S, Quiroz R, et al. Electronic alerts to prevent
droitin sulfate is a contaminant in heparin associated with ad- venous thromboembolism among hospitalized patients. N Engl
verse clinical events. Nat Biotechnol (in press). J Med 2005;352:969-77.
8. Kishimoto TK, Viswanathan K, Ganguly T, et al. Contami- 11. Baroletti S, Munz K, Sonis J, et al. Electronic alerts for hos-
nated heparin associated with adverse clinical events and activa- pitalized high-VTE risk patients not receiving prophylaxis:
tion of the contact system. N Engl J Med 2008;358:2457-67. a cohort study. J Thromb Thrombolysis 2008;25:146-50.
9. Cohen AT, Tapson VF, Bergmann JF, et al. Venous thrombo- Copyright © 2008 Massachusetts Medical Society.
A Small Molecule for a Large Disease
Reed E. Pyeritz, M.D., Ph.D.
Almost three decades ago, Victor McKusick and supporting structures, the notion arose that the
I reviewed Marfan’s syndrome in the Journal and various features of Marfan’s syndrome developed
advised traditional medical and surgical approach- because of “weak” connective tissue. Neat as this
es to management.1 The holy grail at that time was concept was, it did not explain many aspects of
the cause of this autosomal dominant condition, the disease, including overgrowth of tubular
based on the common presumption that under- bones, underdevelopment of muscle and adipose
standing the cause would lead directly to effective tissue, and decreased bone mineral density.
therapy. In fact, life expectancy improved dramat- The creation of mouse models of Marfan’s syn-
ically in subsequent years for all but the most drome, by introducing mutations known to cause
severely affected patients, despite a lack of under- the human disease into the mouse fibrillin gene,
standing of the underlying connective-tissue de- permitted focused investigations of pathogenesis.
fect.2 Dissection of the aorta was and remains the The fibrillins contain several motifs, including
most common cause of death. However, the evo- multiple copies homologous to latent transform-
lution of ever more effective surgical techniques, ing growth factor β (TGF-β) binding protein
applied prophylactically, substantially prolonged (LTBP). The cytokine TGF-β is bound and kept
life.3 Treatment with beta-adrenergic blockade re- inactive by the LTBP complex. Hence, a defect in
tarded the rates of aortic dilatation and dissec- fibrillin structure might be expected to reduce
tion.4,5 Serious complications, including mitral binding and increase the activity of TGF-β. A num-
regurgitation, cardiomyopathy, pneumothorax, de- ber of groups began exploring whether TGF-β
formity of the thoracic cage, myopathy, and di- had a role in Marfan’s syndrome. Dietz and col-
minished visual acuity, still afflict some patients, leagues studied the mouse model and showed
especially infants and children at the most severe that affected tissues had clear evidence of over-
end of the phenotypic continuum.6 Ironically, as expression of TGF-β. The development of cystic
people live longer with Marfan’s syndrome, fea- lungs (which in humans predipose to pneumotho-
tures that previously were unlikely to cause prob- rax), myxomatous mitral-valve leaflets, and aortic
lems, such as lumbosacral dural ectasia, become dilatation are all associated with increased signal-
major issues.7 In addition, new age-dependent fea- ing by TGF-β. Moreover, these common features
tures have been recognized, such as renal and he- of Marfan’s syndrome can be prevented in mice by
patic cysts and biliary stones.8 administering an antibody that binds TGF-β.10‑12
In 1991, the basic defect in Marfan’s syndrome These results dramatically altered the under-
was discovered to be mutation of FBN1, the gene standing of the pathogenesis of this archetypal
that encodes the large extracellular glycoprotein heritable disorder of connective tissue. Notions of
fibrillin-1.9 This discovery raised expectations of therapy quickly switched from “strengthening”
more effective therapies, or even cure. However, the extracellular matrix or replacing the defective
the initial interpretations arising from this dis- fibrillin to modulating signaling through a rea-
covery proved to be inaccurate. Given that the fi- sonably well understood pathway. The idea of de-
brillins are integral components of vascular elas- veloping a small molecule that can be taken oral-
tic fibers, ocular zonules, and other extracellular ly for treatment of the disease gained currency.
n engl j med 358;26 www.nejm.org june 26, 2008 2829