A phase 1b/2 clinical study of marstacimab, targeting human tissue factor pathway inhibitor, in haemophilia

Summary

A phase 1b/2, three‐month study of marstacimab, a human monoclonal antibody targeting tissue factor pathway inhibitor (TFPI), was conducted in participants with haemophilia A or B, with or without inhibitors. Participants assigned to four cohorts received escalating weekly doses based on inhibitor status (without inhibitors: 300 mg, a single 300‐mg loading dose with subsequent 150‐mg doses, or 450 mg; with inhibitors: 300 mg). Safety outcomes were treatment‐emergent adverse events (TEAEs), injection site reactions, clinical and laboratory parameter changes. Efficacy was assessed by annualised bleeding rates (ABRs). Pharmacokinetics and pharmacodynamics (PD) were also evaluated. Among 26 treated participants [haemophilia A without inhibitor, n = 16 (61.5%); haemophilia A with inhibitor, n = 7 (26.9%); haemophilia B, n = 3 (11.5%)], 24 completed the study. Overall, 80.8% experienced TEAEs. ABR during treatment was significantly reduced versus an external on‐demand control group (p < 0.0001) and versus pretreatment ABR (p < 0.0001), with significant reductions observed across all dose cohorts. Marstacimab exposure generally increased in a dose‐related manner, with steady‐state concentration reached by day 57. Changes in pharmacodynamic biomarkers occurred across all dose cohorts. Marstacimab was safe and well tolerated. Clinically meaningful reductions in ABR and treatment‐related changes for all PD biomarkers indicated effective targeting of TFPI. (Clinicaltrials.gov identifier, NCT02974855).

INTRODUCTION

Haemophilia is an X‐linked bleeding disorder caused by deficiencies in clotting factors VIII (FVIII) or IX (FIX). ¹ For patients with severe haemophilia [factor activity levels <1% (<0.01 iu/ml) ² ], intravenous prophylactic clotting factor replacement reduces bleeding and prevents or slows progression of joint disease. ³ , ⁴ , ⁵ However, frequent intravenous infusions lead to treatment non‐adherence and increased treatment burden. ⁶ , ⁷ , ⁸ , ⁹ Limited venous access may require indwelling catheter insertion, which may be complicated by infection or thrombosis. ¹⁰ A subset of patients with haemophilia A or B develop neutralising antibodies (inhibitors) against clotting factor ² , ¹¹ ; non‐factor haemostatic therapies may address these unmet needs. ¹²

Tissue factor pathway inhibitor (TFPI) antagonises early coagulation stages by inhibiting tissue factor–activated coagulation factor VII (FVIIa) and activated factor X. ¹³ , ¹⁴ TFPI is expressed in three isoforms, all sharing the Kunitz‐type inhibitor domain K2 that binds and inhibits activated factor X. Marstacimab (PF‐06741086; Pfizer Inc.), an investigational human immunoglobulin G1 monoclonal antibody, targets the K2 domain and is intended for prophylaxis to prevent or reduce bleeding in patients with haemophilia.

This study evaluated the short‐term safety, tolerability, efficacy, pharmacokinetics (PK) and pharmacodynamics (PD) of weekly subcutaneous (SC) marstacimab administered for up to three months in participants with severe haemophilia A or B, with or without inhibitors, after which participants would be eligible to enrol in a phase 2 long‐term treatment study.

METHODS

This phase 1b/2, open‐label, multicentre study (NCT02974855) was performed at eight centres in North America, South America, Europe and Africa in compliance with the Declaration of Helsinki and Good Clinical Practice guidelines. ¹⁵ , ¹⁶ Institutional review boards and/or independent ethics committees approved the protocol and amendments. All participants provided written informed consent.

Study population

Eligible participants were males aged 18–65 years with haemophilia A or B (≤1% FVIII or FIX activity), with or without inhibitors, previously treated on demand and with at least six acute, non‐surgical, treated bleeding episodes during the six‐month period prior to screening. Participants with inhibitors to FVIII or FIX had to have a titre greater than the upper limit of normal within six months before study day 1 and be receiving a bypass agent for treatment of bleeding episodes. Exclusion criteria included known coronary artery, thrombotic or ischaemic disease; renal, hepatic, haematologic or coagulation laboratory values outside of prespecified limits and inability to substitute treatment with recombinant FVIIa (rFVIIa) for aPCC (complete eligibility criteria, see Appendix S1).

Study design

Four dose cohorts were planned; participants without inhibitors were assigned via an interactive technology system to subcutaneous (SC) administered marstacimab at dose levels of 300 mg once weekly (300‐mg cohort), a single 300‐mg loading dose followed by 150 mg once weekly (300‐mg/150‐mg cohort), 450 mg once weekly (450‐mg cohort), or, for participants with inhibitors, 300 mg once weekly (300‐mg with inhibitors cohort) (Figure 1). All doses of marstacimab were administered at participants' respective study sites. Dose selection was informed by safety and tolerability data from the previous study in healthy participants¹⁷ and by predicted EC90 exposure thresholds for selected PD biomarkers based on results from that study (see Appendix S1). A flat dosing regimen was incorporated because of perceived ease of use and administration and potential benefit in simplifying treatment regimens.

FIGURE 1.

Dose progression scheme. Dose escalation was permitted following review of available safety, pharmacokinetic and pharmacodynamic data at day 29. SC, subcutaneous; QW, once weekly.

Concentration of marstacimab solution for injection was 100 mg/ml. Injection volumes were 1 ml or less, except for the 450‐mg cohort participants, who received three injections of 1.5 ml each to achieve a 450‐mg dose. Treatment began with the 300‐mg cohort; the 300‐mg/150‐mg cohort was a step‐down dose, initiated while safety data from the 300‐mg cohort were maturing to enable dose escalation. Dose escalation and treatment administration in the 450‐mg cohort was permitted following data review for the 300‐mg cohort on day 29. Dosing was stopped if protocol‐specified stopping criteria were met or if available data indicated the current dose was not well tolerated (see Appendix S1).

Weekly study site visits between days 1 and 29 were required, with additional visits on days 2 and 4 for phlebotomy. If a dose within a given cohort was well tolerated through day 29, treatment continued for two subsequent months (days 30–85), with visits on days 30 and 33 for phlebotomy, treatment on days 36, 43, 50, 57, 64, 71 and 78, follow‐up on day 85 and end‐of‐study procedures on day 113. Participants could be followed further if positive for antidrug or neutralising antibodies.

Breakthrough bleeding episodes in non‐inhibitor participants were treated with FVIII or FIX products, with dosing targeted to the lowest level within the local standard of care for achieving haemostasis at the respective bleeding site. For participants with inhibitors, breakthrough bleeding episodes were treated with rFVIIa at doses of approximately 90 μg/kg.

Safety assessments

Safety monitoring included treatment‐emergent adverse events, serious adverse events, injection site reactions, laboratory tests, vital signs, electrocardiograms and physical examinations (see Appendix S1). Adverse event severity was graded per National Institute of Allergy and Infectious Diseases Division of Microbiology and Infectious Diseases Adult Toxicity Tables. ¹⁸

Efficacy assessments

Efficacy was assessed by comparing the treated annualised bleeding rate (ABR) during the study treatment period with the treated ABR from an external control group of 65 individuals who received on‐demand moroctocog alfa (AF‐CC) (ReFacto AF; Pfizer Inc.) or nonacog alfa (BeneFIX; Pfizer Inc.) in previous studies ³ , ¹⁹ , ²⁰ and matched key inclusion/exclusion criteria of the current study (aged 18 to <65 years; factor activity ≤1%). Pfizer internal studies were selected for this comparison group to provide individual participant‐level data that were prospectively collected.

The ABR during the three‐month study treatment period was also compared with the pretreatment ABR calculated retrospectively from the medical record. On‐study ABR was defined as the number of treated bleeding episodes within nine days after last dose/[(last dose date + 9 − first dose date + 1)/365.25] and pretreatment ABR was the number of treated bleeding episodes within six months pre‐enrolment × 2.

Pharmacokinetic and pharmacodynamic assessments

Methodology and timing for PK and PD assessments are summarised in Appendix S1.

Antidrug antibody assessments

Blood samples to provide plasma were collected at prespecified time points, from day 1 to day 113, to assess frequency of antidrug antibody (ADA) and, if positive, neutralising antibody production against marstacimab (see Appendix S1).

Statistical analysis

Safety and ADAs were assessed in all participants who received a dose of marstacimab and efficacy was assessed in the per‐protocol population (all marstacimab‐dosed participants except those with major protocol deviations). The PK and PD populations comprised all participants with samples for at least one PK or PD parameter, respectively.

Annualised bleeding rate was calculated using a negative binomial regression model. The ratio of the ABR and its 80% two‐sided confidence interval (CI) in the marstacimab arm versus that in the external on‐demand control group was generated for each dose cohort. For this early‐stage clinical trial, the on‐treatment ABR with marstacimab would be considered superior to the external on‐demand treatment ABR if the upper limit of the two‐sided 80% CI of the marstacimab/external control group ratio was less than 1 and the ratio estimate was 0.3 or less; if the ratio estimate was above 0.3, marstacimab would be considered not superior to on‐demand treatment. Family‐wise error rate was controlled by testing in a hierarchical order, beginning with the cohort that had the highest pharmacokinetic exposure. Descriptive statistics were used to summarise safety, efficacy, pharmacokinetic and pharmacodynamic data by dose cohort.

RESULTS

Participants

Of 38 patients screened, 27 were enrolled (Figure 2). One participant enrolled in the 300‐mg cohort withdrew consent before treatment. Among 26 treated participants, 24 (92.3%) completed the study and two (7.7%) discontinued because of adverse events. All seven participants in the 300‐mg inhibitor cohort had haemophilia A. One 450‐mg cohort participant had a dose reduction to 300 mg following multiple injection site reactions; this participant was analysed in the 450‐mg cohort. Participants were male; most were aged 18–44 years (Table 1). Three participants had haemophilia B. Most participants had target joints (n = 24/26) and had haemophilic arthropathy (n = 22/26).

FIGURE 2.

Study flow diagram. *One participant excluded because of treatment interruption of more than 30 days caused by a treatment‐emergent adverse event of appendicitis. ^†One participant excluded because of a treatment duration of less than 30 days. AE, adverse event; SC, subcutaneous; QW, once weekly.

TABLE 1.

Participant demographics and baseline disease characteristics

Parameter	Marstacimab dose
	300‐mg cohort (n = 7)	300‐mg/150‐mg cohort (n = 6)	450‐mg cohort (n = 6)	300‐mg with inhibitors cohort (n = 7)	Total (N = 26)
Age, years
Mean ± SD	31.9 ± 8.2	28.7 ± 8.3	41.7 ± 15.9	44.1 ± 9.4	36.7 ± 12.1
Median (minimum, maximum)	28.0 (26, 49)	25.5 (21, 43)	44.5 (19, 58)	40.0 (36, 63)	35.5 (19, 63)
18–44 years, n (%)	6 (85.7)	6 (100.0)	3 (50.0)	4 (57.1)	19 (73.1)
≥45 years, n (%)	1 (14.3)	0	3 (50.0)	3 (42.9)	7 (26.9)
Race, n (%)
White	3 (42.9)	2 (33.3)	6 (100.0)	3 (42.9)	14 (53.8)
Black	4 (57.1)	4 (66.7)	0	4 (57.1)	12 (46.2)
Ethnicity, n (%)
Hispanic or Latino	5 (71.4)	2 (33.3)	3 (50.0)	1 (14.3)	11 (42.3)
Not Hispanic or Latino	2 (28.6)	4 (66.7)	3 (50.0)	6 (85.7)	15 (57.7)
Height, mean ± SD, cm	174.1 ± 3.3	171.9 ± 13.1	179.5 ± 7.3	173.1 ± 7.4	174.6 ± 8.3
Weight, mean ± SD, kg	75.9 ± 13.2	62.3 ± 9.8	84.0 ± 7.1	72.7 ± 16.4	73.8 ± 13.9
BMI, mg/kg, mean ± SD	25.1 ± 4.5	21.0 ± 1.4	26.1 ± 1.1	24.1 ± 4.3	24.1 ± 3.7
Type of haemophilia, n (%)
Haemophilia A	6 (85.7)	5 (83.3)	5 (83.3)	7 (100.0)	23 (88.5)
Haemophilia B	1 (14.3)	1 (16.7)	1 (16.7)	0	3 (11.5)
Target joints, n (%)	7 (100.0)	5 (83.3)	6 (100.0)	6 (85.7)	24 (92.3)
Haemophilia arthropathy, n (%)	7 (100.0)	3 (50.05)	6 (100.0)	6 (85.7)	22 (84.6)
ABR a
Mean ± SD	23.0	14.7	20.3	17.3	18.8
Median (minimum, maximum)	24 (12, 30)	15 (12, 16)	17 (12, 42)	18 (12, 20)	17 (12, 42)

Abbreviations: ABR, annualised bleeding rate; BMI, body mass index.

^a Defined as the number of bleeding episodes within 6 months prior to study enrolment × 2.

In the external control group, median age was 34.2 years (range, 18.7–63.0), 39 controls (60%) had haemophilia A, 26 (40%) had haemophilia B, 57 (87.7%) had haemophilic arthopathy and 57 (87.7%) had target joints.

Safety

Fifty‐six adverse events were reported in 21 (80.8%) participants (Table 2); most were grade 1 or 2 in severity; most common were injection site pain, injection site swelling and hypertension. Four (15.4%) participants experienced a serious adverse event (appendicitis, physical assault, cholelithiasis and tooth socket haemorrhage); none were treatment‐related. No thrombotic events occurred.

TABLE 2.

Summary of treatment‐emergent adverse events

Parameter	Marstacimab dose
	300‐mg cohort (n = 7)	300‐mg/150‐mg cohort (n = 6)	450‐mg cohort (n = 6)	300‐mg with inhibitors cohort (n = 7)	Total (N = 26)
Overview of treatment‐emergent adverse events, n (%)
Any adverse event	7 (100.0)	4 (66.7)	6 (100.0)	4 (57.1)	21 (80.8)
Serious adverse event	1 (14.3)	1 (16.7)	1 (16.7)	1 (14.3)	4 (15.4)
Treatment‐related adverse event	4 (57.1)	4 (66.7)	3 (50.0)	3 (42.9)	14 (53.8)
Dose reduction owing to adverse event	0	0	1 (16.7) a	0	1 (3.8)
Study discontinuation owing to adverse event	0	1 (16.7)	0	1 (14.3)	2 (7.7)
Adverse events in >5% of patients overall, n (%)
Injection site pain	1 (14.3)	1 (16.7)	1 (16.7)	0	3 (11.5)
Injection site swelling	0	0	2 (33.3)	1 (14.3)	3 (11.5)
Hypertension	0	2 (33.3)	1 (16.7)	0	3 (11.5)
Injection site bruising	2 (28.6)	0	0	0	2 (7.7)
Fatigue	1 (14.3)	0	1 (16.7)	0	2 (7.7)
Injection site induration	1 (14.3)	0	1 (16.7)	0	2 (7.7)
Troponin I increased	1 (14.3)	0	0	1 (14.3)	2 (7.7)
Arthralgia	1 (14.3)	0	0	1 (14.3)	2 (7.7)
Prothrombin time prolonged	0	0	1 (16.7)	1 (14.3)	2 (7.7)
Periodontitis	0	1 (16.7)	0	1 (14.3)	2 (7.7)
Influenza	0	1 (16.7)	0	1 (14.3)	2 (7.7)
Headache	0	0	2 (33.3)	0	2 (7.7)

^a One patient had a dose reduction from 450 mg to 300 mg owing to multiple episodes of injection site reactions; the patient was analysed in the 450‐mg cohort.

Twenty‐six treatment‐related adverse events occurred in 14 (53.8%) participants; most common being injection site reactions (eight participants). Treatment‐related grade 3 severity events occurred in four participants: in the 450‐mg cohort, swelling at the injection site occurred in one participant and multiple episodes of pain during injection occurred in another, both in association with injection volumes of 1.5 ml; in the 300‐mg with inhibitors cohort, one participant had generalised pruritus and erythematous rash from days 50 to 113 and injection site erythema, pruritus and swelling from days 50 to 64 (completed study, treatment permanently suspended following day 50 dosing); and one participant had more than a 50% reduction (day 1: 608 mg/dl; day 22: 313 mg/dl) in fibrinogen relative to baseline (but not relative to the screening result) and met prespecified study discontinuation criteria. One participant in the 300‐mg/150‐mg cohort discontinued because of onset of hypertension (considered treatment‐related and met prespecified criteria for dose‐limiting toxicity). Decreases in fibrinogen relative to baseline were observed at all dose levels but were not below the lower limit of normal, did not appear to be dose‐dependent, and were reversible upon treatment discontinuation. Instances of increased cardiac troponin I levels were observed in three participants (one each in the 300‐mg, 300‐mg with inhibitors and 300‐mg/150‐mg cohorts). One 63‐year‐old participant had a treatment‐related grade 1 increase in cardiac troponin (0.10 ng/ml; normal: <0.03 ng/ml) on study day 22 that returned to normal by day 28, without discontinuation of the study drug. A grade 1 increase in cardiac troponin of 0.04 and 0.06 ng/ml, respectively, also occurred in two participants, aged 27 and 43 years; these events were not considered to be treatment‐related and resolved without treatment. None of the cardiac troponin I elevations had corresponding adverse events or electrocardiograph changes indicating increased cardiovascular risk. Prolonged prothrombin time (PT) was observed at a single time point in two participants (one with inhibitors in the 300‐mg cohort and one without inhibitors in the 450‐mg cohort), one of whom had a prolonged PT at baseline. Other adverse events of interest are described in Appendix S1.

Efficacy

The per‐protocol population (n = 24) excluded one participant (300‐mg cohort) because of treatment interruption longer than 30 days owing to appendicitis and another (300‐mg with inhibitors) for treatment duration less than 30 days.

Compared with the external on‐demand control group, mean (80% CI) ABR (negative binomial model) was significantly lower in the on‐study pooled dose cohorts [2.67 (1.83, 3.89) vs 27.62 (24.63, 30.98); ratio (80% CI): 0.10 (0.07 to 0.14); p < 0.0001]. This reduction was significant for each dose cohort (all p ≤ 0.0005; Table 3), with respective ABRs ranging from 0.72 to 4.20. Mean (80% CI) ABR (negative binomial model) also decreased significantly from 18.83 (17.10, 20.74) pretreatment to 2.67 (1.79, 3.97) on‐study [ratio (80% CI): 0.14 (0.09, 0.22), p < 0.0001]. This reduction was significant for each dose cohort (all p ≤ 0.0154; Table 3).

TABLE 3.

Summary of on‐study treatment versus external on‐demand control group and pretreatment annualised bleeding rates

Marstacimab dose	Mean (80% CI) annualised bleeding rate a		Ratio (80% CI)	p value
	On‐study	External on‐demand control group b
300‐mg cohort c	4.20 (2.23, 7.89)		0.15 (0.08, 0.29)	0.0002
300‐mg/150‐mg cohort	1.47 (0.56, 3.90)		0.05 (0.02, 0.14)	0.0001
450‐mg cohort	4.17 (2.21, 7.87)		0.15 (0.08, 0.29)	0.0002
300‐mg with inhibitors cohort			0.03 (0.01, 0.10)	0.0005
Total	2.67 (1.83, 3.89)	27.62 (24.63, 30.98)	0.10 (0.07, 0.14)	<0.0001

Marstacimab dose	Mean (80% CI) annualised bleeding rate a		Ratio (80% CI)	p value
	Pretreatment	On‐study b
300‐mg cohort c	23.00 (19.70, 26.85)	4.19 (2.72, 6.45)	0.18 (0.11, 0.31)	0.0001
300‐mg/150‐mg cohort	14.67 (13.91, 15.47)	1.45 (0.67, 3.15)	0.10 (0.04, 0.22)	0.0002
450‐mg cohort	20.33 (15.76, 26.23)	4.17 (1.99, 8.74)	0.20 (0.09, 0.47)	0.0154
300‐mg with inhibitors cohort	17.33 (15.95, 18.83)	0.73 (0.23, 2.29)	0.04 (0.01, 0.14)	0.0005
Total	18.83 (17.10, 20.74)	2.67 (1.79, 3.97)	0.14 (0.09, 0.22)	<0.0001

Abbreviation: CI, confidence interval.

^a Analysed using a negative binomial model.

^b The external on‐demand control group consisted of 65 participants who received on‐demand treatment with moroctocog alfa or nonacog alfa in previous studies. ³ , ¹⁹ , ²⁰

^c One participant excluded because of treatment interruption of more than 30 days owing to appendicitis.

Pharmacokinetics

Median plasma concentration–time profiles of marstacimab are presented in Figure 3; PK parameters are summarised in Appendix S1. Following administration of a single dose on day 1, maximum concentration (C _max) was reached and median time to peak drug concentration (T _max) values ranged from 69.7 to 71.6 h across dose cohorts. Plasma concentrations of marstacimab increased with dosing up to day 29 in all dose cohorts. On day 29, marstacimab appeared to be absorbed more rapidly, with median T _max values of 22.8–58.5 h across cohorts. Marstacimab exposure [area under the curve (AUC) and C _max] generally increased in a dose‐related manner over the 300‐mg to 450‐mg dose range following both single‐dose and multiple‐dose administration, with steady‐state concentration reached by day 57. Marstacimab concentration–time profiles appeared similar between participants with and without inhibitors following 300‐mg dosing (Figure 3).

FIGURE 3.

Median plasma marstacimab concentrations over time. Levels of marstacimab in plasma were determined at baseline, during the study treatment period (days 1–85), and at follow‐up on day 113. For three of the four study cohorts, the initial dose administered was 300 mg. LD, loading dose; SC, subcutaneous.

Pharmacodynamics

Treatment‐related changes were observed for all PD assessments, across all dose cohorts, and were mostly consistent between participants with and without inhibitors and with haemophilia A and B (Appendix S1). Total TFPI increased in all dose cohorts (Figure 4A), as did peak thrombin generation levels (Figure 4B) and endogenous thrombin generation potential (data not shown). Median absolute peak thrombin generation levels increased within 24 h after the first dose of marstacimab and generally remained steady over time. Treatment‐related shortening of thrombin generation lag time (Figure 4C) occurred in all dose cohorts, as did shortening of dilute PT, which was especially rapid during the first week of treatment (Figure 4D); overall, the greatest shortening occurred in the 450‐mg cohort. Median fibrinogen levels remained within the range 200–400 mg/dl (Figure 5A); prothrombin fragment 1 + 2 levels increased (Figure 5B) in all dose cohorts. Treatment‐related increases in D‐dimer were also observed in all cohorts; however, variability was observed in the 300‐mg/150‐mg and 300‐mg with inhibitors cohorts (Figure 5C).

FIGURE 4.

Summary of pharmacodynamic results. Total tissue factor pathway inhibitor (A); thrombin generation peak (B); thrombin generation lag time (C); and dilute prothrombin time (D). LD, loading dose; SC, subcutaneous.

FIGURE 5.

Coagulation markers of fibrinogen (A), prothrombin fragments 1 + 2 (B), and D‐dimer (C). LD, loading dose; SC, subcutaneous.

Antidrug antibodies

No participants had positive ADA samples (titre ≥ 1.53) at baseline. After treatment initiation, positive ADA results were detected for three (11.5%) participants, with titres ranging from 1.72 to 2.89. None tested positive for neutralising antibodies, and no impact on safety, PK, or PD was observed in ADA‐positive participants.

DISCUSSION

Once‐weekly, SC marstacimab treatment was generally safe and well tolerated at all dose levels. Significant reductions in ABR occurred at all dose levels, relative to a prespecified external control group and relative to baseline pretreatment ABRs. Treatment‐related changes indicative of haemostatic pharmacology were observed for PD biomarkers. Results were similar between participants with and without inhibitors.

No thrombotic events or serious treatment‐related adverse events occurred. Injection site reactions were the most common adverse event, occurring in nearly one third of participants; however, most were grade 1 severity. Two participants in the 450‐mg cohort had grade 3 severity injection site reactions, likely related to the associated increase in injection volume (1.5 ml). No clinically significant changes in safety laboratory parameters were observed. Reversible fibrinogen reduction occurred during treatment at all dose levels, but values never fell below the lower limit of normal. Elevations in D‐dimer, a prespecified PD end‐point, were observed; however, these were in the absence of clinically significant changes in platelet count, prothrombin time/international normalized ratio, or activated partial thromboplastin time, or changes in levels of fibrinogen that would pose safety concerns. No thromboembolic events were reported and thrombin generation assay showed no evidence of excessive thrombin generation. Single isolated instances of low‐grade cardiac troponin I elevation in each of three participants occurred in the absence of electrocardiogram changes, adverse events, or other information that would confer the strong pretest likelihood required to diagnose a myocardial infarction. ²¹

This study did not incorporate a concurrent internal control group. However, the prespecified external control group was compiled from subject‐level data for individuals matched for key inclusion/exclusion criteria who received on‐demand treatment during prospective studies and reflects the natural history of haemophilia, which remains unchanged in individuals receiving on‐demand treatment. ³ , ¹⁹ , ²⁰ , ²² This approach incorporated recommendations for use of historic control groups, accounted for important/prognostic baseline characteristics that meet eligibility criteria in the current study and resulted in an external control group that is similar to the study population with regard to age, proportion with haemophilic arthropathy and proportion with target joints. It enabled a model‐based ABR that is a conservative reflection of the proportion of ABRs reported during on‐demand treatment in adolescent/adult patient populations, ⁵ , ²³ , ²⁴ , ²⁵ and augmented the approach used for other non‐factor treatments of comparing ABR during three months of study treatment with pretreatment ABR, as retrospectively determined. ²⁶ The determination of efficacy relative to an external control group is supported by the significant reduction from the pretreatment ABR observed at all marstacimab dose levels evaluated. Given the similar range and/or overlapping responses of coagulation biomarkers (Figures 4 and 5), comparable reductions in ABR observed at all dose levels was not unexpected. Although indirect comparisons between studies of prophylactic interventions are limited by differences in trial designs, treatment regimens and study populations, the respective on‐study ABR of 2.67 (0.65–4.22 across cohorts) is consistent with that reported in prior studies in participants receiving prophylactic factor replacement ²⁷ , ²⁸ , ²⁹ and with the ABR in recent phase 2 studies of another daily SC TFPI K2 inhibitor in development for treatment of patients with haemophilia, concizumab (Novo Nordisk). ³⁰

The previously reported half‐life of marstacimab after a single dose was approximately three days, and responses of key coagulation biomarkers at a 300‐mg dose level persisted for at least 168 h, ¹⁷ providing rationale for this study's once‐weekly regimen. Marstacimab has a longer half‐life (several days ¹⁷ ) than TFPI, explaining the observed increases in total TFPI levels, as the half‐life of bound TFPI, that is, total TFPI, reflected the half‐life of marstacimab rather than its own. ³¹

Other limitations of this study include its small sample size, with only three participants with haemophilia B (none with inhibitors), a three‐month duration of treatment and evaluation of only one dose level in patients with inhibitors. A companion phase 2 long‐term treatment study extends initial safety and efficacy assessments and includes provisions to evaluate the 300‐mg/150‐mg dose level in participants with inhibitor (NCT03363321). Other anti‐TFPI agents in development (e.g., concizumab and befovacimab) have been associated with instances of thrombosis, ³² , ³³ , ³⁴ and thrombotic safety remains a key consideration for non‐clotting factor haemostatic treatments. ³⁵ , ³⁶ , ³⁷ A more definitive assessment of thrombotic risk of marstacimab in patients with haemophilia awaits completion of an ongoing phase 3 study (NCT03938792). Given similar efficacy at the 150‐mg and 450‐mg weekly dose levels, the 300‐mg loading dose followed by 150 mg once weekly, being evaluated in the ongoing phase 3 study, provides opportunity to establish a wide therapeutic index. ³⁸

Safety, efficacy, PK and PD profiles of marstacimab support further evaluation of once‐weekly SC dosing in patients with severe haemophilia, with or without inhibitors. Clinically meaningful reductions in ABRs were observed across all dose levels, consistent with treatment‐related changes observed for all PD biomarkers, indicating effective targeting of the TFPI pathway.

AUTHOR CONTRIBUTIONS

Steven Arkin contributed to the study design. Johnny N. Mahlangu, Silva Zupančić Šalek, Michael Wang and Lisa N. Boggio served as study investigators and provided patients or study materials. Johnny N. Mahlangu participated in the collection and assembly of data. Johnny N. Mahlangu and Pengling Sun participated in data analysis. Johnny N. Mahlangu, Michael Wang and Steven Arkin contributed to data interpretation. Johnny N. Mahlangu, Silva Zupančić Šalek, Michael Wang and Steven Arkin participated in the preparation of the manuscript. All authors participated in the critical review and revision of this manuscript and provided approval of the manuscript for submission.

FUNDING INFORMATION

This study was sponsored by Pfizer Inc.

CONFLICT OF INTERESTS

Johnny N. Mahlangu has received research grants from Bayer, Biogen, Biomarin, CSL Behring, Novo Nordisk, Sobi, Roche and Unique; is a member of scientific advisory committees for Amgen, Bayer, Biotest, Biogen, Baxalta, CSL Behring, Catalyst Biosciences, Novo Nordisk, Roche and Spark; is a member of speakers' bureaus for Alnylam, Bayer, Biotest, Biogen, Novo Nordisk, Pfizer, Sobi, Shire, Roche, ISTH and World Federation of Haemophilia. Andrzej Mital, JLL, Juan Cristobal Morales and Daniel R. Malan have no conflicts of interest to disclose. Silva Zupančić Šalek has received honoraria from Novo Nordisk, Shire and Sobi, and has served on an advisory board for Roche. Michael Wang has received research funding from Bayer, Biomarin, Bioverativ/Sanofi, CSL Behring, Genentech, Novo Nordisk, Octapharma, Pfizer, Shire/Takeda and uniQure, and has served on a board or advisory committee for Bayer, BioMarin, Bioverativ/Sanofi, Catalyst Biosciences, CSL Behring, Genentech, Novo Nordisk and Shire/Takeda. Lisa N. Boggio has received research funding from Biomarin, CSL Behring, Genentech, Novo Nordisk, Pfizer and Sanofi, and has served on an advisory committee for Bayer, CSL Behring, Genentech, Novo Nordisk and Octapharma. Inga Hegemann has received honoraria from and served as a consultant for Bayer, CSL Behring, Novo Nordisk, Octapharma, Shire, Sobi and uniQure. Matthew Cardinal, Tong Zhu, Pengling Sun and Steven Arkin are employees of Pfizer and may own stock/options in the company.

DATA AVAILAIBILITY STATEMENT

Upon request, and subject to review, Pfizer will provide the data that support the findings of this study. Subject to certain criteria, conditions and exceptions, Pfizer may also provide access to the related individual de‐identified participant data. See https://www.pfizer.com/science/clinical‐trials/trial‐data‐and‐results for more information.

Supporting information

Appendix S1

Mahlangu JN, Lamas JL, Morales JC, Malan DR, Šalek SZ, Wang M, et al. A phase 1b/2 clinical study of marstacimab, targeting human tissue factor pathway inhibitor, in haemophilia. Br J Haematol. 2023;200(2):229–239. 10.1111/bjh.18420