Eli Lilly and Company is seeking approval of drotrecogin alfa (activated) for the treatment of adult and pediatric patients with severe sepsis. Information in this document provides an overview of data contained in the Biologics License Application submitted to the United States Food and Drug Administration on 25 January 2001. The proposed indication statement for drotrecogin alfa (activated) reads: "Drotrecogin alfa (activated) is indicated for the treatment of adult and pediatric patients with sepsis associated with acute organ dysfunction (severe sepsis). Treatment with drotrecogin alfa (activated) reduces mortality in patients with severe sepsis." Drotrecogin alfa (activated) will be recommended for use as adjunctive therapy to best standard of care in the treatment of patients with severe sepsis.

An estimated 750,000 episodes of severe sepsis occur annually in the Unites States (Linde-Zwirble et al. 1999). Severe sepsis is now the thirteenth most common cause of death in the United States and is among the most common causes of death in the non-coronary intensive care unit (Balk 2000). Few clinical syndromes are associated with such a rapidly progressive clinical course as severe sepsis: between 3 and 5 of every 10 patients die within 4 weeks of diagnosis (Rangel-Frausto et al. 1995; Natanson et al. 1998). This extremely high mortality rate persists despite best standard of care including administration of appropriate antibiotic therapy, adequate source control of infection, and support for failing organs.

The development of severe sepsis is associated with a generalized inflammatory and procoagulant response, as manifested by increased levels of pro-inflammatory cytokines and markers of thrombin generation. These inflammatory and procoagulant host responses are tightly and intricately linked. In spite of adequate antimicrobial therapy and removal of the source of infection, an intensive systemic host response can lead to extensive endothelial damage, fibrin deposition in the microvasculature, organ hypoperfusion, parenchymal cell dysfunction, multiple organ dysfunction, and death.

Activated Protein C, an endogenous protein that promotes fibrinolysis and inhibits thrombosis and inflammation, is an important modulator of the coagulation and inflammation associated with severe sepsis. Activated Protein C is converted from its inactive precursor, Protein C, by thrombin coupled to thrombomodulin (Esmon 1989). Reduced levels of Protein C are found in the majority of patients with sepsis and are associated with an increased risk of death. In addition to low plasma levels of Protein C, the conversion of Protein C to Activated Protein C may be impaired during sepsis as a result of the down-regulation of thrombomodulin by inflammatory cytokines.

The above observations lead to the development of recombinant human Activated Protein C [drotrecogin alfa (activated)] as a therapy for patients with severe sepsis. Data supporting the safety and efficacy of drotrecogin alfa (activated) for the treatment of adult patients with severe sepsis was derived from a single, multi-country, placebo-controlled Phase 3 study F1K-MC-EVAD (n=1690 patients) with supporting data from a single, multi-country, placebo-controlled Phase 2 study (n=131 patients). Analysis of data from the Phase 2 study F1K-MC-EVAA indicated that drotrecogin alfa (activated), administered as a 24 mg/kg/hr constant rate infusion for 96 hours, reduced markers of coagulopathy (D-dimer) and inflammation (IL-6) compared to placebo. Analysis of data from the Phase 3 study F1K-MC-EVAD indicated that drotrecogin alfa (activated), administered as a 24 mg/kg/hr constant rate infusion for 96 hours, significantly reduced 28-day all cause mortality in patients with severe sepsis compared to placebo (relative risk = 0.8057; p=0.0054). At Study Day 28, the observed mortality rates were 24.71% for the drotrecogin alfa (activated) group and 30.83% for the placebo group. These data represent an unprecedented breakthrough: in the pivotal Phase 3 study F1K-MC-EVAD, an additional 6 lives were saved for every 100 patients treated with drotrecogin alfa (activated). Consistent with its antithrombotic and profibrinolytic effects, the administration of drotrecogin alfa (activated) was associated with an increase in the percent of patients experiencing a bleeding event reported as a serious adverse event (3.5% vs. 2.0%; p=0.06). Serious bleeding events frequently resulted from injury to a blood vessel (traumatic or iatrogenic) or following instrumentation of a highly vascular organ, such as the kidney or lung. There were no other safety concerns associated with the administration of drotrecogin alfa (activated) to patients with severe sepsis.

Data supporting the use of drotrecogin alfa (activated) in the pediatric population (newborn to 17 years) was derived from a single, open-label Phase 1B study which provided pharmacokinetic, pharmacodynamic and safety data for drotrecogin alfa (activated) in pediatric patients with severe sepsis (N=83 patients). Although limited by patient number, analysis of data from the Phase 1B study indicated that the pharmacokinetics, pharmacodynamics (as assessed by serial measures of D-dimer), and safety profile of drotrecogin alfa (activated) were similar between pediatric and adult patients with severe sepsis. As the Phase 1B study in pediatric patients with severe sepsis was an open-label study, the efficacy of drotrecogin alfa (activated) for the treatment of severe sepsis in pediatric patients must be extrapolated from the well-controlled Phase 3 study conducted in adults.

Additional safety and effectiveness data on the use of drotrecogin alfa (activated) for the treatment of severe sepsis in pediatric and adult patients are currently being obtained in a global, open-label study (estimated enrollment at completion of the study = 2,500 patients). The ongoing open-label study employs similar inclusion criteria, exclusion criteria, and dosing regimen as were used in the pivotal Phase 3 study.

The data contained within the BLA demonstrate that drotrecogin alfa (activated) has a favorable benefit risk profile. In the population studied, one additional life was saved for every 16 patients treated with drotrecogin alfa (activated).

ALT alanine aminotransferase
ANOVA analysis of variance
APACHE II Acute Physiology and Chronic Health Evaluation II
APC Activated Protein C
APTT activated partial thromboplastin time
D APTTmax maximum change from baseline in APTT response
AST aspartate aminotransferase
BLA Biologics License Application
BT bleeding time
CI confidence interval
Clp plasma clearance
Css plasma concentration at steady-state
CVVHD continuous venovenous hemodialysis
DIC disseminated intravascular coagulation
DSMB Data Safety Monitoring Board
EPCR endothelial Protein C receptor
FDA Food and Drug Administration
F1.2 prothrombin fragment F1.2
HSA human serum albumin
ICH International Conference on Harmonisation
ICU intensive care unit
IL-6 interleukin-6
LLN lower limit of normal
LOCF last observation carried forward
M males
max maximum value measured
min minimum value measured
MODS multiple organ dysfunction syndrome
PAI-1 plasminogen activator inhibitor-1
PC Protein C
PT prothrombin time
D PTmax maximum change from baseline in PT response
SD standard deviation
SIRS systemic inflammatory response syndrome
SOFA Sequential Organ Failure Assessment
t½ half-life in plasma
t½a half-life in plasma during the initial, rapid distribution phase
t½b half-life in plasma during the second, slower elimination phase
tlast sampling time of the last quantifiable plasma concentration
TAFI thrombin activatable fibrinolysis inhibitor
TATc thrombin-antithrombin complex
ULN upper limit of normal
Vss volume of distribution at steady-state

Approximately 750,000 cases of sepsis associated with acute organ dysfunction (severe sepsis) occur annually in the United States (Angus et al. 2001). The mortality rates associated with severe sepsis in the United States range from 28% to 50% and have remained essentially unchanged for several decades (Natanson et al. 1998). Each year, 215,000 deaths are associated with severe sepsis; deaths after acute myocardial infarction occur at approximately an equal rate (Natanson et al. 1998; Angus et al. 2001; Murphy 2000).

In 1992 a Consensus Panel of the American College of Chest Physicians and the Society of Critical Care Medicine met to address several issues related to the concepts and terminology associated with the host response to infection (Bone et al. 1992; Wenzel et al. 1996). The central hypothesis for the proposals developed by this Consensus Panel was that the clinical manifestations associated with serious infections were primarily related to the host response to the infection rather than the underlying infectious agent. Furthermore, similar clinical manifestations could also be induced by non-infectious processes that also led to excessive activation of the inflammatory response pathways. The term systemic inflammatory response syndrome (SIRS) was introduced to describe the constellation of clinical manifestations of the systemic inflammatory response to injury. SIRS is defined as the presence of two or more objective signs of systemic inflammation in the absence of evidence of an infectious disease. SIRS may result from a variety of pathologic insults, such as pancreatitis, burns, and trauma.

Objective signs of systemic inflammation include fever, hypothermia, tachycardia, tachypnea, and neutrophilia or neutropenia. These manifestations result from the release of inflammatory mediators (cytokines, eicosanoids, proteases, kinins, etc.) in response to pathologic insults. Excessive release of these inflammatory mediators may result in the development of diffuse capillary injury, parenchymal cell dysfunction, intravascular coagulation with microvascular thrombosis, and multiple organ dysfunction. The term multiple organ dysfunction syndrome (MODS) is used to describe the development of organ dysfunction in patients with SIRS.

When two or more objective signs of systemic inflammation occur in the presence of a known or suspected infection, the term sepsis is used rather than SIRS. Sepsis is usually the result of a serious bacterial infection, but may occur in response to other pathogens such as fungi, viruses, and parasites. When an excessive inflammatory response leads to organ dysfunction in patients with sepsis, the term severe sepsis is used.

The inflammatory response is normally regulated by a network of endogenous anti-inflammatory mediators, coagulation inhibitors, and fibrinolytic components. These regulatory systems maintain a state of homeostasis in blood flow, endothelial cell function, and organ function. The progression from sepsis to severe sepsis is associated with loss of this homeostasis. Of note, progression to severe sepsis is associated with marked activation of the coagulation system with depletion of endogenous regulatory components. In addition, after initial activation, the fibrinolytic system becomes relatively inhibited due to an increase in plasminogen activator inhibitor-1 (PAI-1). These changes lead to an unbalanced coagulation system that favors thrombin generation and fibrin clot formation.

The historical approach to the treatment of sepsis has included antimicrobial therapy and supportive therapies such as vasopressors and ventilatory management. However, even with optimal antimicrobial and supportive therapy, the mortality rate among patients with severe sepsis remained high at 28% to 50%. The failure of conventional therapy to further improve the clinical outcome for these patients led to the evaluation of therapies directed at restoration of homeostasis in the host response to the infection. Following discoveries on the biochemical mechanisms of the inflammatory response in the early 1980s, therapies designed to inhibit the inflammatory response were evaluated in patients with sepsis. Targets of these therapies included pro-inflammatory cytokines, endotoxin, complement, adhesion molecules, and pro-inflammatory metabolic pathways. These clinical trials demonstrated little, if any, positive effect on clinical outcome measured as all-cause mortality, usually at 28 days following start of therapy. The failure of these measures to improve survival has led to exploration of alternative mechanisms to modulate other pathways involved in the pathophysiology of severe sepsis.

The development of severe sepsis is associated with a generalized inflammatory and procoagulant response, as manifested by increased levels of pro-inflammatory cytokines and markers of thrombin generation. These inflammatory and procoagulant host responses are tightly and intricately linked. In spite of adequate antimicrobial therapy and removal of the source of infection, an intensive systemic host response can lead to extensive endothelial damage, fibrin deposition in the microvasculature, organ hypoperfusion, parenchymal cell dysfunction, multiple organ dysfunction, and death.

The association between activation of coagulation and the development of septic shock was delineated more than 30 years ago (Corrigan et al. 1968). Of note, activation of coagulation was found to be independent of the type of infectious microorganism; gram-positive bacteria, gram-negative bacteria, and parasites were all shown to be capable of triggering this response. Subsequently, abnormalities in fibrinolysis, in addition to coagulation, were frequently documented in patients with sepsis (McGilvary and Rotstein 1998; Vervloet et al. 1998; Levi and ten Cate 1999; van Gorp et al. 1999). Our evolving understanding of the host response at the molecular level in the last three decades has continued to shed new light on the links between infection, inflammation, and hemostasis.

The link between sepsis and activation of coagulation is perhaps most overt in the subset of patients presenting with bruising and skin lesions referred to as purpura fulminans. Skin biopsies from these patients reveal extensive microvascular thrombosis. Of note, the histopathology of these lesions is very similar to the findings in patients with neonatal purpura fulminans, a disorder associated with a congenital absence of Protein C or Protein S, two proteins that are necessary for the regulation of thrombin formation.

Host monocytes and macrophages, in response to infectious pathogens/products, generate and release inflammatory cytokines such as tumor necrosis factor-a (TNF-a ), interleukin-1 (IL-1), and interleukin-6 (IL-6) (Parrillo 1993). Although these early response cytokines play a critical role in host defense by attracting activated neutrophils to the site of infection, the entry of these cytokines and pathogen products into the systemic circulation can bring about widespread activation of coagulation and suppression of fibrinolysis (van der Poll et al. 1990; van Deventer et al. 1990; Carvalho and Freeman 1994; Hinshaw 1996). The infectious agents and inflammatory cytokines activate coagulation by stimulating the surface expression of tissue factor on monocytes and the endothelium. The exposure of tissue factor to circulating blood initiates coagulation activation that leads to the generation of thrombin and fibrin deposition (clot). For example, in an animal endotoxemia model, it was demonstrated that microthrombi developed in the hepatic microcirculation within 5 minutes of endotoxin challenge (Asaka et al. 1996). If endotoxin exposure continued, multiple fibrin clots developed and resulted in focal areas of hypoperfusion, tissue necrosis, and development of multiple organ dysfunction.

Thrombin is a potent serine protease that has a number of functions, including multiple pro-inflammatory properties (Esmon 2000a). The increased thrombin generation resulting from coagulation activation induced by infectious agents and inflammatory cytokines initiates a vicious cycle by further intensifying the host inflammatory and coagulation response.

As part of the host’s attempt to interrupt this vicious cycle and re-establish homeostasis, anti-inflammatory cytokines are released (Bone et al. 1997; Dinarello 1997; Antonelli 1999; van der Poll and van Deventer 1999; Calandra and Heumann 2000; Cavaillon and Adib-Conquy 2000). On the side of hemostasis, the host’s endogenous fibrinolytic system and anticoagulant systems are brought into action to try to counter the excessive coagulation activation. Key components of the fibrinolytic system include: tissue plasminogen activator (tPA), which initiates the generation of plasmin; plasminogen, which, when converted to the active enzyme plasmin, is responsible for lysis of fibrin clots; PAI-1, a potent inhibitor of tPA; and thrombin activatable fibrinolysis inhibitor (TAFI) which, when activated by thrombin, suppresses the activity of plasmin (Bazjar 2000). The host’s fibrinolytic system is impaired during sepsis by the inflammatory mediators that stimulate the release of PAI-1 from platelets and the endothelium. Consequently, PAI-1 levels are elevated and tPA activity is suppressed in patients with sepsis (Suffredini et al. 1989; Vervloet et al. 1998).

The Protein C pathway is one of three major anticoagulant systems involved in the regulation of thrombin formation (Figure 1.1). Central to the Protein C pathway is the vitamin K-dependent factor Protein C. Protein C is the inactive precursor (zymogen) of the serine protease Activated Protein C. Protein C is converted to Activated Protein C by thrombin in complex with an endothelial surface receptor called thrombomodulin (Esmon 1989). The activation of Protein C is further augmented by another endothelial surface protein, endothelial Protein C receptor (EPCR) (Esmon et al. 1999). Activated Protein C inactivates Factors Va and VIIIa, two key factors in the formation of thrombin. Factor Va accelerates the activation of thrombin by Factor Xa, whereas Factor VIIIa accelerates the activation of Factor X by Factor IXa. Inactivation of Factors Va and VIIIa by Activated Protein C thus limits the generation of thrombin and is a potent antithrombotic mechanism.

Many studies have shown that Protein C is depleted in both adult and pediatric patients with sepsis and that there is an inverse correlation between the level of Protein C and mortality and morbidity outcomes in these patients (Fourrier et al. 1992; Lorente at al. 1993; Powars et al. 1993; Boldt et al. 2000; Fisher and Yan 2000). From the study reported by Mesters and coworkers, the decrease in plasma Protein C levels preceded the onset of the clinical symptoms of severe sepsis and septic shock by a median of 12 hours, indicating that depletion of Protein C occurs early in the disease course (Mesters et al. 2000). This early decrease in Protein C levels in the pathogenesis of sepsis is further supported by the prevalence (>90%) of acquired Protein C deficiency in patients with severe sepsis (Hartman et al. 1998; Yan et al. 2001). The depletion of Protein C in patients is most probably due to a combination of several mechanisms. Protein C is susceptible to degradation by neutrophil elastase, which is released during sepsis (Philapitsch and Schwartz 1993). The continuous and rapid conversion of Protein C to Activated Protein C in sepsis can lead to depletion of the plasma pool of Protein C. The biosynthesis of Protein C to replenish the circulating pool, a process that is dependent on the liver, may be inadequate due to hepatic dysfunction or acquired vitamin K deficiency.

In addition to the decreased plasma concentration of Protein C, the conversion of Protein C to Activated Protein C may be impaired in patients with severe sepsis. In vitro studies have shown that endotoxin and inflammatory cytokines, such as TNF-a , can down-regulate the endothelial surface thrombomodulin either by decreasing synthesis or by increasing degradation of thrombomodulin (Moore et al. 1987; Moore et el. 1989; Lentz et al. 1991). Endothelial surface thrombomodulin is also cleaved by neutrophil elastase and released into the circulation as soluble thrombomodulin (MacGregor et al. 1997). An elevation of circulating soluble EPCR has been demonstrated in both an experimental mouse sepsis model (Gu et al. 2000) and in patients with sepsis (Kurosawa et al. 1998). Most recently, endothelial surface thrombomodulin and EPCR were shown to be reduced in skin biopsy samples from 21/21 patients and 17/21 patients, respectively, with meningococcal septicemia (Faust et al. 2000). Thus in patients with severe sepsis, even though there is an elevation of thrombin generation, the conversion of Protein C to Activated Protein C may be limited by the combination of decreased plasma concentrations of Protein C and a decrease in the concentration of endothelial surface thrombomodulin and EPCR. This conclusion is supported by data on the levels of Activated Protein C in an animal model of bacteremia (Taylor et al. 2000b) and data from patients with severe sepsis (Mesters et al. 2000), suggesting that the rise in Activated Protein C in sepsis is transient and does not parallel with the continuous rise in thrombin levels.

In addition to impairment of the Protein C system, the plasma concentration of antithrombin is reduced in patients with severe sepsis. The decrease in antithrombin also appears to occur early in the disease process. Overall, the data suggest that hemostasis is unbalanced in patients with severe sepsis, with an increase in coagulation activation and thrombin generation and impairment of the hemostatic regulatory systems. This imbalance contributes to enhancement of the inflammatory response, microvascular hypoperfusion, organ dysfunction, and the high mortality in patients with severe sepsis.

Activated Protein C has been shown to have profibrinolytic activity in an animal model (Jackson et al. 1998). The profibrinolytic activity of Activated Protein C is an indirect effect mediated by three possible molecular mechanisms. First, Activated Protein C inhibits PAI-1. This inhibition involves the formation of a stable complex between PAI-1 and Activated Protein C. PAI-1 trapped in this complex is not capable of inhibiting tPA. Consequently, high levels of Activated Protein C may result in less inhibition of tPA because of the competition for PAI-1. Second, in vitro data suggest that Activated Protein C may inhibit the release of PAI-1 from endothelial cells, again resulting in less inhibition of tPA. Third, activation of TAFI is dependent on high concentrations of thrombin. Inhibition of thrombin generation by Activated Protein C thus limits the activation of TAFI (Bazjar et al. 1996).

The proposed anti-inflammatory activities of Activated Protein C, a more recent discovery, are based mostly on in vitro data (Grinnell and Yan 1998). Since thrombin has been shown to have multiple pro-inflammatory activities, limiting thrombin generation by Activated Protein C would have indirect anti-inflammatory effects (Esmon 2001). In addition, Activated Protein C has been show to directly interact with endothelial cells and leukocyte membranes in combination with EPCR (Esmon 2000b). This Activated Protein C:EPCR membrane complex has been shown to alter cell responses. Activated Protein C reduces NFk B nuclear translocation resulting in reduction of cytokine synthetic rates (Esmon 2000c). Activated Protein C decreases the expression of leukocyte adhesion molecules on the surface of endothelial cells, resulting in decreased interaction between these cells, a key step in the movement of leukocytes out of the vessels and into the tissues. Anti-apoptotic pathways are activated following exposure of cells to Activated Protein C; activation of these pathways is associated with decreased apoptosis in response to the apoptotic agent staurosporin (Joyce et al. 2001). The molecular mechanism for these effects is still uncertain, but may involve translocation of Activated Protein C to the nucleus of inflammatory cells, affecting the transcriptional activity of acute phase response genes (White et al. 2000).

In view of the role of unbalanced activation of coagulation in the pathophysiology of severe sepsis, a molecule such as Activated Protein C, with its multiple proposed mechanisms of action, is an attractive candidate for clinical evaluation. Activated Protein C may help break the vicious cycle of the host response, leading to improvement in organ function and survival.

Experiments with the baboon model of bacteremia provide additional support for the potential role of administration of Activated Protein C to patients with severe sepsis (Taylor et al. 1987; Taylor et al. 2000a). In one series of experiments, all control animals receiving a lethal dose of bacteria died of sepsis-related complications, whereas all of the animals that received the same dose of bacteria in conjunction with administration of Activated Protein C survived. Administration of Activated Protein C was also associated with decreased thrombin generation and amelioration of the coagulopathy in the septic animals, as monitored by a variety of biomarkers.

Subsequent experiments explored the role of EPCR in the host response to sepsis. Animals were given either a control antibody or an antibody that blocked the interaction between EPCR and Protein C, followed by a sublethal dose of bacteria. All animals receiving the control (non-inhibitory) antibody survived, whereas all animals that received the inhibitory antibody died of sepsis-related complications. The fatal outcome in the animals receiving the inhibitory antibody was presumed to be due to decreased activation of Protein C. In another series of experiments, baboons were treated with the control antibody or the inhibitory antibody and then infused with thrombin to stimulate activation of Protein C. Administration of the blocking antibody to EPCR inhibited Protein C activation by 88% (Taylor et al. 2001). Thus, inhibition of EPCR function leads to decreased activation of Protein C and poorer outcomes in this model of sepsis.

Taken together, these animal model data provide additional evidence supporting the hypothesis that Activated Protein C may be effective at reducing mortality among patients with severe sepsis. In addition, these data demonstrate the importance of an intact endothelial system for generation of Activated Protein C in vivo. The observation that endothelial expression of both EPCR and thrombomodulin are suppressed in patients with severe sepsis suggests that the endogenous system for generating Activated Protein C is diminished. Administration of the activated form of Protein C bypasses this defect associated with sepsis because the active enzyme does not rely on the host’s endothelial thrombomodulin or EPCR for conversion. For this reason, the activated form of Protein C is preferred over the precursor (zymogen) form for the treatment of patients with severe sepsis.

The plasma concentration of Activated Protein C (1 to 2 ng/mL) and zymogen Protein C (approximately 4000 ng/mL) in healthy humans is too low to support large scale production of Activated Protein C from the human blood supply. Consequently, recombinant DNA technology is used by the sponsor to produce human Activated Protein C. The proposed generic name for the recombinant human Activated Protein C produced by the sponsor’s process is drotrecogin alfa (activated).

Human Protein C was cloned and a suitable expression system developed (Beckmann et al. 1985; Grinnell et al. 1987; Yan et al. 1990). Because of the very complex structure of human Protein C requiring four different types of post-translational modifications for full biological activity, recombinant human Activated Protein C has to be produced using a mammalian cell line (Grinnell et al. 1987; Yan et al. 1990). The recombinant molecule is identical to the plasma-derived human Activated Protein C in its protein sequence. It is distinguished only by differences in the carbohydrate portion of the molecule (Yan et al. 1993).

The results of the baboon studies indicated a potential mortality benefit of Activated Protein C for sepsis (Section 1.6). In support of the clinical program, an extensive series of pharmacodynamic, pharmacokinetic, and toxicology studies were conducted with drotrecogin alfa (activated). The only adverse event reported in these studies was bleeding. These bleeding events were generally categorized as nonserious and involved bleeding or oozing at venipuncture or surgical sites. Spontaneous, nontraumatic hemorrhage was uncommon. The preclinical toxicology studies of drotrecogin alfa (activated) provided supporting data for the initial clinical development program of drotrecogin alfa (activated) to a maximum infusion rate of 50 m g/kg/hr.

The mortality of severe sepsis remains at 30% to 50% in spite of advances in antimicrobial therapies and life support modalities. Severe sepsis is manifested with a systemic host response of inflammation and coagulopathy that leads to endothelial injury, fibrin deposition, hypoperfusion, multi-organ dysfunction, and death. The multiple mechanisms of action of Activated Protein C (antithrombotic, profibrinolytic, and anti-inflammatory) and the safety profile of drotrecogin alfa (activated) from preclinical toxicology studies support the evaluation of this compound as a treatment of severe sepsis in improving survival. The learning and experience gained from past clinical trials in sepsis in the last two decades have been invaluable in shaping the design of the clinical trial protocols for this program.

Drotrecogin alfa (activated) has been studied in a variety of patient populations, including healthy subjects and patients with end-stage renal disease (Phase 1), as well as adult and pediatric patients with severe sepsis (Phase 1B and Phase 2/3). The information presented in this section contains an overview of each of these studies including study design, objectives, patient population, dose duration, and results.

Eight Phase 1 studies were completed and included in the BLA (Table 2.1). The primary objective of the initial seven studies of the Phase 1 program was to evaluate the safety of drotrecogin alfa (activated) in healthy male and female subjects who received doses up to 48 µg/kg/hr for durations up to 24 hours. An additional study evaluated safety in male and female subjects with end-stage renal disease receiving either hemodialysis or peritoneal dialysis.

Study F1K-MC-EVAO is an open-label, dose-escalation study that is currently ongoing at 12 sites in the United States and the United Kingdom. This study is designed to investigate the pharmacokinetics and safety of drotrecogin alfa (activated) administered to pediatric patients with severe sepsis. Additionally, the pharmacodynamics of drotrecogin alfa (activated), as assessed by changes in plasma coagulation parameters (D-dimer concentration, Protein C activity level, and antithrombin activity level), will be investigated. In Part 1, infusions of drotrecogin alfa (activated) administered at 6, 12, 24, and 36 m g/kg/hr for 6 hours once daily over a 4-day treatment period in three age groups (newborn to <1 year, ³ 1 year to <8 years, and ³ 8 years to <18 years) were evaluated for an appropriate dose for Part 2. In Part 2, patients received a 96-hour infusion at
24 m g/kg/hr (based on Part 1 results) (Table 2.2).

Drotrecogin alfa (activated) has been evaluated in the treatment of severe sepsis in two randomized, double-blind, placebo-controlled, multicenter clinical studies (Table 2.3).

Study F1K-MC-EVAA was a Phase 2 dose-ranging study conducted at 40 investigative sites in the United States and Canada. The primary objectives of this study were to evaluate the safety of administration of drotrecogin alfa (activated) as a function of infusion rate and infusion duration and the degree to which the coagulation abnormalities of severe sepsis were affected by the administration of drotrecogin alfa (activated) as a function of infusion rate and infusion duration; and to determine an effective infusion rate and infusion duration of drotrecogin alfa (activated) administration based on its ability to alter the coagulation abnormalities of severe sepsis. The study was divided into two stages. During Stage 1, patients were randomly assigned to receive placebo or drotrecogin alfa (activated) at an infusion rate of 12, 18, 24, or 30 m g/kg/hr administered for a 48-hour infusion duration. During Stage 2, patients were randomly assigned to receive placebo or drotrecogin alfa (activated) at an infusion rate of 12, 18, or 24 m g/kg/hr administered for a 96-hour infusion duration. All patients were followed for 28 days or until death. Of the 135 patients with severe sepsis enrolled in the study, 131 received study drug: 4 patients withdrew from the study before receiving any study drug, 90 patients received drotrecogin alfa (activated), and 41 patients received placebo.

Study F1K-MC-EVAD was a Phase 3 pivotal efficacy and safety trial conducted at 164 investigative sites in 11 countries (Australia, Brazil, Belgium, Canada, France, Germany, Netherlands, New Zealand, Spain, South Africa, and the United States) in 1728 randomly assigned patients with severe sepsis. Of these patients, 1690 received study drug (850 [drotrecogin alfa (activated)], 840 placebo). This study compared infusion of drotrecogin alfa (activated) administered at 24 m g/kg/hr with infusion of placebo for 96 hours. The primary objective of the study was to demonstrate a reduction in 28-day all-cause mortality with treatment of drotrecogin alfa (activated) compared with placebo.

Phase 1 Studies. The clinical pharmacology and pharmacokinetics of drotrecogin alfa (activated) were evaluated in eight Phase 1 studies. These studies, comprising
112 unique subjects, are summarized in an integrated fashion.

The primary objective of the Phase 1 program was to evaluate the safety of drotrecogin alfa (activated) over a range of doses up to 48 m g/kg/hr for 24 hours. Phase 1 studies were conducted in healthy males and females and in: (a) healthy females with low estrogen levels, (b) subjects undergoing hemodialysis or peritoneal dialysis as therapy for end-stage renal disease, and (c) healthy males and females pretreated with aspirin. Of the 172 subjects entered in Phase 1 studies, 27 were not exposed to drotrecogin alfa (activated) and 33 participated in two to four study protocols.

The safety profile of drotrecogin alfa (activated) in the Phase 1 studies was evaluated by measuring the effect of drotrecogin alfa (activated) on the following physiological parameters:

• Bedside whole blood activated partial thromboplastin time (APTT)
• Template bleeding time (BT)
• Bedside whole blood prothrombin time (PT)
• Factor V and Factor VIII
• Complete blood count
• Platelet count
• Occult fecal blood and occult urine blood
• Serum chemistries
• Electrolytes
• Urinalysis
• Spontaneous and solicited adverse events
• Anti-APC antibody formation.

Conventional noncompartmental and compartmental methods were used to calculate the following primary pharmacokinetic parameters in Phase 1 studies:

• Plasma concentration at steady-state (Css)
• Plasma clearance (Clp)
• Volume of distribution at steady-state (Vss)
• Half-life in plasma (t1/2)

Plasma concentrations of APC were measured using an immunocapture-amidolytic activity assay specific for APC (Gruber and Griffin 1992). As Activated Protein C is inhibited by several plasma protease inhibitors (Heeb et al. 1989; Marlar et al. 1993; Scully et al. 1993), blood samples were collected into citrate tubes containing benzamidine, a reversible inhibitor of Activated Protein C that prevents inhibition by the plasma protease inhibitors. Benzamidine inhibition was reversed after immunocapture of Activated Protein C and removal of the plasma protease inhibitors, thus restoring the amidolytic activity of Activated Protein C. The assay has a validated range of 1 to
10 ng/mL at the low end and 100 to 200 ng/mL at the high end. Samples with values above the upper limit of quantitation were diluted and re-analyzed.

Pharmacokinetic data of drotrecogin alfa (activated) were derived using the immunocapture assay method, and thus reflect the plasma concentration of active enzyme. The assay method does not distinguish between drotrecogin alfa (activated) and endogenous Activated Protein C. However, the concentration of endogenous Activated Protein C is generally below the lower limit of quantitation. The assay method also does not provide data on the plasma clearance of Activated Protein C-protease inhibitor complexes or Activated Protein C/drotrecogin alfa (activated) metabolites.

Throughout the descriptions of Phase 1 results, the term "normal healthy subjects" identifies a subset of Phase 1 subjects that excludes estrogen-deficient women, subjects requiring hemodialysis, subjects requiring peritoneal dialysis, and bolus-dosed subjects. These subjects were omitted from the "normal healthy" database because estimates of Clp in these subjects were statistically significantly different from those in normal healthy subjects in at least one Phase 1 study (estrogen-deficient women and subjects requiring hemodialysis), because the population studied was not healthy (subjects requiring hemodialysis or peritoneal dialysis), or because parameter estimates were either not calculable or unreliable because steady-state was not attained (bolus-dosed subjects).

Estimates of Css and Clp were robust and were not greatly affected by excluding specific subjects. However, excluding estrogen-deficient women, subjects requiring hemodialysis, subjects requiring peritoneal dialysis, and bolus-dosed subjects preserved the integrity of the database attributed to the normal healthy population, and ensured that parameter estimates ascribed to that population were based on data from subjects who truly were "normal" and "healthy."

Half-Life of Drotrecogin Alfa (Activated) in Healthy Subjects. Elimination of drotrecogin alfa (activated) was biphasic and rapid. Plasma t1/2 was 0.913 ± 0.679 hr (N=251 doses) based on all Phase 1 subjects and 0.693 ± 0.411 hr (N=78 doses) based on the more homogenous subset of normal healthy subjects who received drotrecogin alfa (activated) at a dose of 25.1 ± 2.0 m g/kg/hr (Table 3.1). Because of assay sensitivity at lower doses, these half-life estimates were hybrids of t1/2 measured during the initial and terminal phases, and depended on infusion rate and duration. However, the estimates in all subjects and in normal healthy subjects are comparable, which indicates that the overall estimate of t1/2 obtained from Phase 1 subjects is robust and is not substantially affected by choice of subject subpopulation.

Because t1/2 is so short, Css is reached rapidly after starting a constant rate infusion. Elimination of drotrecogin alfa (activated) is biphasic, with a rapid initial phase (t1/2a ) of 13 minutes and a slower second phase (t1/2b ) of 1.63 hours. The short t1/2a of 13 minutes accounts for 79% of the area under the plasma concentration curve, and governs the initial rapid accrual of plasma APC concentrations toward steady-state levels. The longer t1/2b of 1.6 hours controls the time it takes to get from 90% of steady-state to 100% of steady-state, and governs the time it takes to eliminate the final 21% of drotrecogin alfa (activated) infused during treatment. Approximately 75%, 90%, and 97% of Css will be reached within 40 minutes, 1.8 hours, and 4.5 hours, respectively, of starting an infusion. Likewise, those are the fractions of drug that will be eliminated within those times after stopping an infusion. Therefore, the short t1/2 of drotrecogin alfa (activated) confers rapid attainment to steady-state during infusion and then rapid elimination after infusion.

Steady-State Plasma Concentration of Drotrecogin Alfa (Activated) in Healthy Subjects. Plasma concentrations of endogenous APC in healthy subjects were usually below detection limits and did not significantly influence the pharmacokinetics of drotrecogin alfa (activated). Steady-state plasma concentrations were proportional to infusion rate in all studies in which infusion rate varied (Figure 3.1). Steady-state plasma concentrations did not depend on infusion duration when infusions at one rate were given to the same subjects for 6 hours and 24 hours (Studies F1K-LC-GUAC and F1K-LC-GUAD).

Based on integrated results from the Phase 1 studies (Table 3.1), Css at an infusion rate of 24 m g/kg/hr is expected to be 71.2 ng/mL based on all subjects, 68.9 ng/mL based on normal healthy subjects, and 72.4 ng/mL based on normal healthy subjects infused at 25.1 ± 2.0 m g/kg/hr.

Clearance of Drotrecogin Alfa (Activated) in Healthy Subjects. Pharmacokinetic data produced during drotrecogin alfa (activated) development indicate the following:

• Clp of drotrecogin alfa (activated) in healthy subjects does not depend on infusion rate or infusion duration.
• Clp of drotrecogin alfa (activated) in healthy subjects tends to increase with increasing body weight, thus justifying body weight-normalized dosing in healthy subjects.
• The overall estimate of Clp in healthy subjects is robust, in spite of various subject characteristics that influenced estimates of drotrecogin alfa (activated) Clp in specific Phase 1 studies.

Based on integrated results from the Phase 1 studies (Table 3.1), estimated Clp (mean ±  SD) was 27.0 ± 8.4 L/hr (N=251 doses) based on all doses, 28.1 ± 8.6 L/hr (N=190 doses) based on normal healthy subjects, and 26.0 ± 6.8 L/hr (N=78 doses) based on normal healthy subjects infused at 25.1 ± 2.0 m g/kg/hr.

Volume of Distribution of Drotrecogin Alfa (Activated) in Healthy Subjects. The Vss of drotrecogin alfa (activated) is small, which is consistent with the high molecular weight of drotrecogin alfa (activated) and the presumed effect of that bulk on its ability to penetrate membranes. The Vss of approximately 16 to 20 L in normal healthy subjects is comparable to that of extracellular volume.

Activated Partial Thromboplastin Time. APTT was the predominant pharmacodynamic parameter assessed during the Phase 1 studies. Drotrecogin alfa (activated) is rapidly neutralized by endogenous plasma protease inhibitors; therefore, the determination of APTT and PT were measured by an assay performed at the patient’s bedside using whole blood. The assay was performed within two minutes of the sample being obtained from the subject. Use of whole blood minimized the time necessary to prepare the sample for measurement of APTT and PT. Conduct of the assay at the bedside eliminated the time necessary to transport the sample to the hospital laboratory.

APTT correlated strongly with APC concentration in all individual studies. This relationship is best exemplified by Study F1K-LC-GUAD, in which men and women were infused with drotrecogin alfa (activated) at a mean rate of 12.8, 25.4, 38.0, and 49.9 m g/kg/hr for 6 hours and 24 hours. During these infusions, the time course of the pharmacodynamic response to drotrecogin alfa (activated), expressed as percent change in APTT from baseline (%D APTT), exactly paralleled the time course of drotrecogin alfa (activated) in plasma (Figure 3.2), and showed no evidence of hysteresis. As expected from this relationship, a plot of %D APTT versus plasma APC concentration is linear (Figure 3.3).

Consistent with the rapid decline of APC plasma concentrations after the end of infusion, the APTT at the highest drotrecogin alfa (activated) infusion rate and duration of 48 m g/kg/hr for 24 hours returned to within 20% of its predose levels within 5 hours. APTT returned to within 10% of its predose levels within 4 hours after a 24-hour infusion at 24 m g/kg/hr. These times to return to baseline reflect the time it takes for drotrecogin alfa (activated) activity to effectively dissipate.

Prothrombin Time. PT was a secondary pharmacodynamic parameter monitored during Phase 1 studies. PT correlated weakly but statistically significantly with APC concentration in all individual studies. As exemplified by Study F1K-LC-GUAD, the maximal change in PT from baseline (D PTmax) depended on infusion rate (p<0.001) and, at the higher doses, also depended on infusion duration (p=0.002). Although the observed changes in PT were statistically significant, the difference was less than
3 seconds. In contrast to the proportional response of APTT to Css, PT changed little over a 4-fold range of drotrecogin alfa (activated) infusion rates.

Bleeding Time. In Study F1K-LC-GUAD, there was an effect of drotrecogin alfa (activated) on bleeding time that was dependent on estrogen status (p=0.028). The absolute difference was only 1 to 1.5 minutes over the entire dose range tested in Phase 1 studies.

Platelet Function. In Study F1K-LC-GUAD, platelet function in response to arachidonic acid, adenosine diphosphate, and ristocetin was independent of infusion rate, infusion duration, and estrogen status. The lack of effect of drotrecogin alfa (activated) on platelet function was a consistent finding in all studies in which platelet function was assessed.

Factor V and Factor VIII. In Study F1K-LC-GUAD, Factor V levels decreased significantly during infusion in a manner dependent on infusion duration (p=0.038). Factor V levels did not depend on estrogen status.

Factor VIII levels declined during infusions. The decline depended on infusion duration (p=0.010), but not on estrogen status.

Prolonged administration of drotrecogin alfa (activated) was associated with a statistically significant decline in Factor V and, to a lesser extent, Factor VIII levels. However, both Factor V and Factor VIII levels remained within the normal range and this effect is probably of minor clinical significance.

Serum Electrolytes, Chemistry and Hematology Panels, and Urinalysis. Out-of-range levels from serum electrolyte panels, chemistry and hematology panels, and urinalysis panels were infrequent and appeared to be distributed randomly across the range of drotrecogin alfa (activated) infusion rates.

Safety Findings in the Phase 1 Studies. In these studies, safety was assessed by analyzing adverse events, serious adverse events, laboratory data, and vital sign data. These analyses indicated no toxicities associated with the administration of drotrecogin alfa (activated) that precluded further clinical investigation of drotrecogin alfa (activated) as a therapeutic for patients with severe sepsis.

Mild headache and ecchymosis were the most frequent adverse events reported by subjects in the Phase 1 studies. The incidence of headache depended on infusion rate (dose), but not infusion duration. The etiology of headache is unclear. Most headache resolved with administration of acetaminophen. There was no neurological sequelae or need to discontinue drug as a result of headache. Ecchymosis was reported most often at the site of venipuncture. Additionally, there was a higher than expected occurrence of occult blood in the urine in healthy subjects. Neither of these phenomena were associated with infusion rates (dose) but both of these findings are very likely to be the result of the antithrombotic and profibrinolytic properties of drotrecogin alfa (activated).

In the Phase 1 studies, only 1 subject developed a problem that was reported as three serious adverse events (dysuria, hematuria, and neoplasm). The events occurred during the outpatient waiting period between drotrecogin alfa (activated) infusions (first infusion of 6 hours and second infusion of 24 hours). Subject 2041, a 72-year-old male, developed hematuria 8 days after completion of the 6-hour infusion. His coagulation parameters were normal when he was released to outpatient follow-up. The subject was re-admitted to the clinical pharmacology facility, which met the serious event criteria by extending the subject’s hospitalization and making a diagnosis of cancer. He was found to have a filling defect in his right renal calyx. A follow-up cystoscopy revealed two small tumors in the bladder and a mass in the right renal calyx consistent with transitional cell carcinoma. The subject discontinued from the study and did not receive a second infusion. The subject was transferred into the care of an urologist at the Indiana University Hospitals and then to Veterans Administration for surgery and follow-up. The investigator judged these events to be a typical presentation for transitional cell carcinoma and unrelated to drotrecogin alfa (activated) administration.

Immunogenicity. See Section 10.

• In normal healthy subjects, the pharmacokinetics of drotrecogin alfa (activated) were linear through infusion rates of 48 m g/kg/hr. Vss was small, and elimination was biphasic with a rapid initial phase (t1/2a = 13 minutes) and a slower second phase (t1/2b  = 1.6 hours). Because of these short half-lives, 90% of the steady-state plasma concentration is reached within 2 hours of starting a constant rate infusion, and 90% is eliminated within 2 hours of stopping the infusion.
• The time course of pharmacodynamic response to drotrecogin alfa (activated) in healthy subjects, expressed as change from baseline whole blood APTT (D APTT), paralleled the time course of Activated Protein C concentration in plasma. The change from baseline whole blood prothrombin time (D PT) also increased with increasing plasma APC concentration, but the magnitude of change in D PTmax was very small. Factor V and Factor VIII levels decreased with increasing infusion duration. Estrogen status did not affect the relationship between APTT or PT and plasma APC concentration. There was no evidence of hysteresis in any of the pharmacodynamic parameters affected by APC.
• Mild headache and ecchymosis were the most frequent adverse events reported in healthy subjects.

These pharmacokinetic and safety profiles supported continued development of drotrecogin alfa (activated) as a potential therapy for the treatment of severe sepsis.

The Phase 2 study F1K-MC-EVAA was a randomized, double-blind, placebo-controlled study designed to investigate the safety and pharmacokinetics of drotrecogin alfa (activated) in patients with severe sepsis and to determine an effective infusion rate and duration for use in subsequent studies. An effective infusion rate and infusion duration were determined by evaluating the effect of drotrecogin alfa (activated) administration on the coagulation abnormalities associated with severe sepsis. The primary markers of coagulation measured in the study were D-dimer level, fibrinogen level, and platelet count, and the primary marker of inflammation measured was IL-6. The protocol-specified primary analysis was based on the treatment to which the patient was randomly assigned irrespective of the actual dose received. Supplemental analyses were performed based on drug exposure (m g/kg) and the steady-state concentration achieved (ng/mL).

A statistically significant dose-dependent decline in D-dimer and IL-6 levels was observed in drotrecogin alfa (activated) patients compared with placebo patients. This dose-dependent effect was evident regardless of the type of analysis conducted (ie, by treatment assignment, by drug exposure, or by steady-state concentration achieved). An infusion rate of 24 m g/kg/hr for 96 hours was associated with the largest reduction in D-dimer and IL-6 levels. No safety concerns were identified that would limit the infusion rate or duration. This infusion rate and duration were recommended for use in the Phase 3 pivotal trial (Study F1K-MC-EVAD).

Eligible patients were male or female, 18 years or older with severe sepsis who met criteria for systemic inflammatory response syndrome and associated organ failure. Patients must have met inclusion criteria within a 24-hour time period. From the time a patient met inclusion criteria, an additional 36 hours were allowed for the investigator to obtain informed consent, complete randomization, and initiate the study drug infusion.

This study was conducted in two sequential steps designated as Stage 1 and Stage 2. Both Stage 1 and Stage 2 were randomized, double-blind, placebo-controlled, dose-ranging studies of drotrecogin alfa (activated) or placebo administered as a continuous intravenous infusion over a fixed interval of 48 hours (Stage 1) or 96 hours (Stage 2).

Stage 1. The safety, pharmacokinetics, and pharmacodynamics of a 48-hour infusion of drotrecogin alfa (activated) were evaluated using a dose-escalation scheme with an initial infusion rate of 12 m g/kg/hr. Patients were randomly assigned to drotrecogin alfa (activated) or placebo treatment in a 2:1 ratio. Five dosing groups were formed. After the first dosing group had received an infusion rate of 12 m g/kg/hr, an unblinded Data Monitoring Board was convened to review the available safety, pharmacokinetic, and pharmacodynamic data from this dosing group. Based on these data, the Data Monitoring Board determined the infusion rate to be received by the next dosing group.

Stage 2. A 96-hour infusion was evaluated using a dose-escalation scheme similar to that used in Stage 1. Patients were randomly assigned to drotrecogin alfa (activated) or placebo treatment in a 3:1 ratio. Four dosing groups were formed. The initial infusion rate was determined by the Data Monitoring Board, based on Stage 1 results. As in Stage 1, the unblinded Data Monitoring Board was convened to review the available safety, pharmacokinetic, and pharmacodynamic data from the first dosing group and to determine the infusion rate to be received by the next dosing group.

In both stages, bedside whole blood APTT testing was performed 4 hours after the initiation of the study drug infusion and every 24 hours for the duration of the infusion. The assay was performed within two minutes of the sample being obtained from the subject. Use of whole blood minimized the time necessary to prepare the sample for measurement of APTT and PT. Conduct of the assay at the bedside eliminated the time necessary to transport the sample to the hospital laboratory.

If a patient’s bedside whole blood APTT was ³ 95 seconds, then the patient’s infusion rate was reduced by 25% of the original dose and whole blood APTT was retested in 4 hours. After retesting, if the whole blood APTT was <95 seconds, the current dose was maintained; if the whole blood APTT was ³ 95 seconds, the dose was again reduced by 25% of the original dose. This process continued until the patient’s whole blood APTT was <95 seconds. The decision to reduce the infusion rate based on the bedside whole blood APTT was prospectively defined and based on standard practices for heparin use.

Figure 4.1 shows the study design and treatment groups for Study F1K-MC-EVAA.

Statistical Methods. The primary population of interest consisted of all enrolled patients who received study drug infusion [drotrecogin alfa (activated) or placebo] for any length of time. For analysis, patients were allocated to the treatment group to which they were randomly assigned.

D-dimer and IL-6 levels are presented as percent change from baseline summary statistics at each postbaseline time point. At each time point, analysis of variance (ANOVA) based on ranked data was used to draw statistical conclusions. In addition to the primary analyses, monotonic dose response analyses were performed using the percent change from baseline to end of infusion data for all patients. The last-observation-carried-forward (LOCF) method was chosen as the primary imputation method due to the impact of death and administrative sample handling resulting in missing data. Two-sided p-values £ 0.05 are noted as statistically significant with no adjustment for the multiple comparisons performed.

The impact of drotrecogin alfa (activated) administration on coagulation and inflammation markers was also assessed by allocating patients to treatment groups according to total drotrecogin alfa (activated) drug exposure. The drotrecogin alfa (activated) patient population was segmented into exposure quartiles within each study stage. The total drug exposure for each patient was calculated as the sum of the dose received (m g/kg/hr) at every time point during infusion, thus yielding a total exposure expressed in m g/kg. Patients who died prior to the end of their assigned infusion period had the dose received at the time of death carried forward for the duration of their assigned infusion to eliminate the impact of death on the exposure calculation.

For each primary coagulation and inflammation marker, graphics portraying a correlation analysis of the percent change from baseline to end of infusion marker results with patient predicted APC steady-state concentrations are presented. A two-sided p-value based on a test for zero correlation using Spearman’s rank correlation is calculated.

In Study F1K-MC-EVAA, 135 patients were entered and randomly assigned to drotrecogin alfa (activated) or placebo treatment. Four patients withdrew from the study before receiving study drug: 3 patients did not meet entry criteria and 1 patient withdrew consent. Of the 131 patients who received study drug, 90 received drotrecogin alfa (activated) (Stage 1: 46 patients; Stage 2: 44 patients) and 41 received placebo (Stage 1: 26 patients; Stage 2: 15 patients).

Table 4.1 contains a summary of treatment assignments and the number of drotrecogin alfa (activated) patients who did not receive a full dose of study drug or received a change in dose. Patients were defined as receiving a full dose of study drug if they received a study drug infusion of at least 47 hours during Stage 1 or at least 95 hours during Stage 2 with no change in their initial infusion rate. Patients were defined as having a change in dose if their infusion rate was reduced or increased for dose error reasons, or was reduced because of an elevated bedside whole blood APTT.

Six of the 12 patients who received drotrecogin alfa (activated) 30 m g/kg/hr for 48 hours had a change in infusion rate because of a prolonged bedside whole blood APTT. Also, 6 patients in the 30 m g/kg/hr had maximum Activated Protein C steady-state concentrations greater than 250 ng/mL. Although there were no safety concerns noted in this treatment group, an infusion rate of 30 m g/kg/hr was not evaluated during Stage 2 (96-hour infusion) based on these results.

Table 4.2 contains a summary of the demographic characteristics of patients who received study drug.