Clinical Pharmacokinetics

https://doi.org/10.1007/s40262-019-00751-7

REVIEW ARTICLE

Revisiting the Pharmacology of Unfractionated Heparin

Abdallah Derbalah1   · Stephen Duffull1   · Fiona Newall2,3 · Katie Moynihan4,5   · Hesham Al‑Sallami1 

© Springer Nature Switzerland AG 2019

Abstract
Unfractionated heparin (UFH) is a commonly used anticoagulant therapy for the acute treatment and prevention of thrombosis.
Its short duration of action, reversibility of effect by protamine sulfate, and extensive clinical experience are some of the advan-
tages that support its use. However, the choice of dose and dosing regimen of UFH remains challenging for several reasons.
First, UFH has a narrow therapeutic window and wide variability in the dose–response relationship. Second, its pharmacody-
namic (PD) properties are difficult to characterise owing to the complex multidimensional mechanisms of interaction with the
haemostatic system. Third, the complex heterogeneous chemical composition of UFH precludes precise characterisation of
its pharmacokinetic (PK) properties. This review provides a comprehensive mechanistic approach to the interaction of UFH
with the haemostatic system. The effect of chemical structure on its PK and PD properties is quantitatively described, and a
framework for characterisation of the dose–response relationship of UFH for the purpose of dose optimisation is proposed.

1 Introduction
Key Points 
Despite widespread use of unfractionated heparin (UFH) in
Unfractionated heparin (UFH) is a heterogeneous the treatment and prevention of thrombosis, optimal dosing
mixture of species with anticoagulant activity produced remains a challenge. Non-linearity and wide variability in the
through multiple interactions with the haemostatic sys- dose–response relationship, as well as the narrow therapeutic
tem. Therefore, characterising its pharmacokinetic (PK) window, render accurate prediction of ideal doses difficult.
and pharmacodynamic (PD) properties is difficult. This necessitates frequently repeated measures of various
Because there are no assays to quantify UFH concentra- coagulation response biomarkers to guide dose adjustments.
tion in plasma, no PK analysis of UFH has been reported To achieve successful dose optimisation, all significant
to date. Instead, the PK nature of UFH has been imputed sources of variability in dose response need to be identified
from markers of PD activity. and quantified.
Historically, the pharmacokinetic (PK) and pharmacody-
This review comprehensively and quantitatively namic (PD) properties of UFH have been difficult to char-
describes a full model of UFH interaction with the hae- acterise. This is made more challenging because UFH is a
mostatic system, as far as is known from the literature. heterogeneous mixture of chemicals with subcomponents
having different PD effects. To date, no study has reported
* Abdallah Derbalah
on measuring and performing a PK analysis of UFH. Current
[email protected] literature in this area has imputed the PK nature of UFH from
a biomarker of its PD properties, i.e. using a kinetic-PD (KPD)
1
School of Pharmacy, University of Otago, Dunedin, approach (see Jacqmin et al. [1] and Gabrielsson et al. [2] for
New Zealand
a description of these methods); hence, the true underlying
2
Department of Nursing, The University of Melbourne, PK properties remain unknown. This is further challenged as
Parkville, VIC, Australia
measurement of its PD response is distorted by the variability
3
Department of Paediatrics, The Royal Children’s Hospital, inherent in the haemostatic system that it perturbs [3].
The University of Melbourne, Parkville, VIC, Australia
Several population pharmacological models have been
4
Department of Cardiology, Boston Children’s Hospital, constructed to describe the dose–response relationship of
Boston, MA, USA
UFH [4–7]; however, these KPD models are empirical (data
5
Department of Paediatrics, Harvard Medical School, Boston, driven) and therefore may not reflect the diversity of the
MA, USA

Vol.:(0123456789)
 A. Derbalah et al.

biological system being studied. It is also difficult to identify coagulation system is its ability to amplify the insult signal,
sources of variability in model parameters since these are resulting in rapid fibrin generation at the site of injury. How-
composite of both PK and PD. As a result, dose–response ever, clotting should be controlled to prevent occlusion of the
relationship predictions under conditions other than those whole vessel or extension to non-injured regions, resulting
used to fit the model to the data are problematic [8]. Mecha- in serious thrombotic complications. Antithrombin III (AT)
nistic description of the PK and PD of UFH is thus needed and heparin cofactor II (HCII) are two key anticoagulant
as a basis for identifying, quantifying, and accounting for mechanisms contributing to blood fluidity, providing 70%
the sources of variability in the dose–response relationship of the total clotting enzyme inhibiting capacity of plasma
of UFH and allow for models that can accurately quantify [11]. Because these inhibitors represent the main targets for
dosing requirements. UFH-mediated anticoagulant effects, their biological activity
This review revisits the pharmacology of UFH in order to is discussed in greater detail.
provide a more comprehensive quantitative and mechanistic
review of the PK and PD properties of UFH. We begin with a 2.1 Antithrombin III
brief overview of the haemostatic system, particularly focus-
ing on components modulated by UFH. The next section AT is a 55 kDa member of the protein superfamily of ser-
describes important structural features of UFH that influence pins, sharing many structural and biochemical properties.
its pharmacologic properties. We then explore the interac- AT was initially identified in 1905 as a factor decreasing
tion of UFH with the haemostatic system and the influence the clotting activity of blood [12]. The AT-dependent anti-
of various structural features on that interaction. Finally, the coagulant activity of UFH was subsequently described in
various limitations associated with current assays used to 1939 [13]; however, it was not until 1946 that an assay to
quantify the activity of UFH in plasma are discussed. We measure AT activity became available, enabling the study of
note that since no concentration measurements have been AT in clinical research and practice [14, 15]. An immense
made and analysed at the time of this review, we are not able amount of research ensued and has contributed to our cur-
to explore the foundational PK properties of UFH, and all rent understanding of the activity of this key anticoagulant.
PK characteristics are either imputed based on the properties
of the mixture or its KPD relationships. 2.1.1 Biochemical Basis of Antithrombin (AT) Activity

AT exerts its anticoagulant activity through the formation

2 The Haemostatic System of a stoichiometric (1:1) complex (Eq. 1), thereby inactivat-
ing several key activated clotting factors in the coagulation
The haemostatic system maintains an intricate balance network. The covalent nature of the complex formed means
between blood fluidity (ability of blood to freely flow within it is stable against reducing and denaturing agents [16].
vessels) and coagulability (the ability to reduce blood flow
in localised regions of vessels). For such a balance to exist, kAT−Fa
AT + Fa ⟶ AT:Fa. (1)
both procoagulant and anticoagulant components need to be
tightly regulated. This is achieved through complex interac- In this schematic equation, Fa is used to represent any
tion of cells, coagulation factors, cofactors, and mediators number of activated factors. The primary target for inactiva-
involving a series of positive and negative feedback and tion by AT is IIa, and AT is the major plasma inhibitor of IIa.
feedforward interactions. As a product and key modulator of the coagulation network,
The coagulation network [9] is responsible for regulation tight regulation of IIa is essential for balanced haemostasis.
of the formation and subsequent degradation of the fibrin IIa clearance from blood is mediated either by binding to
component of clots. It is represented as clotting factors endothelial cell receptors with subsequent internalisation,
circulating in inactive form that become readily activated or by formation of an inactive IIa inhibitor complex that is
when the system is stimulated. Traditionally, coagulation rapidly cleared from blood [17].
activation and propagation was theoretically divided into two Factor Xa is another key product of the coagulation net-
pathways—the intrinsic and extrinsic pathways. The extrin- work common pathway that is inactivated by AT. Factor Xa,
sic pathway is stimulated by exposure to tissue factor (TF), together with activated factor V (Va), form a complex that
while the intrinsic pathway is triggered when clotting factor binds the membrane phospholipids of activated platelets and
XII comes in contact with a negatively charged surface. The accounts for the majority of prothrombin activation into IIa,
two pathways converge to a common pathway at factor X and termed the prothrombinase complex. IIa is also generated
conclude by the key reaction of the thrombin (IIa) catalysis outside of the prothrombinase complex but at a much slower
of fibrinogen to fibrin and its subsequent crosslinking, the rate [9]. AT, via the schematic in Eq. 1, can also inactivate
backbone of a blood clot [10]. An important feature of the factors IXa, XIa, XIIa, and VIIa:TF [18].
Pharmacology of Unfractionated Heparin

The rate of clotting factor inactivation by AT can be activity against IIa and does not inhibit other clotting factors.
described by second-order kinetics, with typical values The mechanism of HCII-mediated IIa inhibition is similar to
shown in Table 1. Differences in values may be attributed that of AT [38]. HCII may play an important physiological
to testing under diverse reaction conditions, variable reagent role in neutralising thrombin following activation by der-
potency, or different protein molar masses used in computa- matan sulfate (the most abundant glycosaminoglycan found
tion of the reaction rate constants. For example, it was found in the extracellular matrix of both arterial and venous walls,
that fibrinogen decreases the second-order rate constant of and is a major contributor to the non-thrombogenic proper-
the AT:IIa complex formation, possibly through a competi- ties of blood vessels [39]). Heterozygous HCII deficiency is
tive mechanism, while platelets increase that constant [19]. a mild risk factor for thrombotic diseases, with no definite
homozygous deficiency cases reported in the literature [40].
2.1.2 Pharmacokinetics (PK) of AT HCII forms a 1:1 stoichiometric inactive complex with
IIa (Eq.  2). As with AT, the reaction follows second-
The normal plasma concentration of AT in healthy sub- order kinetics, with values reported from 1.5 × 10 5 to
jects is 2–3 µM (150–200 mg/L) [27], with a half-life of 6 × 105 M−1 min−1, as measured in a pure system [20, 41].
approximately 3 days [28]. The plasma concentration of AT The normal plasma concentration of HCII is 0.92–1.24 µM
in infants is 30–40% less than that of adults and may not (61–82 mg/L) [42]. The half-life of 125I-labelled HCII in
reach adult values until 3 months of age [29]. The volume of healthy volunteers was measured as 2.5 days [42]. Inactive
distribution of AT is 100–115 mL/kg, indicating significant HCII–IIa complexes are rapidly cleared from blood by the
extravascular distribution [30–32]. AT protease complexes liver, as is the case with AT inhibitory complexes [18].
are recognised and cleared by the liver [33]. UFH decreases
the half-life of AT by approximately 25% [28, 31], which kHCII−IIa
HCII + IIa ⟶ HCII:IIa. (2)
contributes to a gradual decrease of AT concentration during
prolonged or high-dose UFH therapy [34].

2.2 Heparin Cofactor II 3 Unfractionated Heparin (UFH)

HCII is another plasma serpin protease inhibitor that is Heparin is a naturally occurring glycosaminoglycan pro-
approximately 30% homologous to AT. It was first described duced by mast cells and basophiles. It was discovered by
in 1967 and was identified as having a similar activity to McLean in 1916 [43] and was employed clinically in 1935
AT but slower electrophoretic mobility [35]. Immunologic as an anticoagulant. Recent studies have demonstrated addi-
assays to measure HCII concentration are relatively simple tional antineoplastic, anti-inflammatory and anti-infective
and provide sensitive and specific measurements [36]. By properties of heparin [44]. The physiological function of
comparison, amidolytic assays are relatively more complex, endogenous heparin is unclear, however its exclusive intra-
requiring an initial step for removal of AT from plasma, cellular presence suggests that it may not have an antico-
however they provide more information about the physio- agulant role [45]. UFH is the least processed form of the
logic activity of HCII [37]. Unlike AT, HCII exhibits specific naturally occurring heparin and is obtained by purification

Table 1  Second-order rate constant kAT−Fa for AT clotting factor complex formation reactions
Clotting factor kAT−Fa ­(M−1 min−1) Reaction conditions References

IIa 4 × 105 Ca2+ (5 mmol/L), phospholipids (20 µg/mL), prothrombin (0.2 mg/mL) [19]

6.3 × 105 Ca2+ + ADP-activated platelets (2 × 105/L)
1.36 × 105 Ca2+, phospholipids, prothrombin + fibrinogen (9 µM)
2 × 106 Pure system [20]
3.4 × 105 Pure system [21]
8.7 × 103 Pure system [22]
Xa 9 × 104 AT-depleted plasma spiked with various AT concentrations [23]
1.8 × 105 Pure system [24]
2.3 × 103 Pure system [22]
IXa 2.9 × 104 Pure system [24]
XIa 3 × 104 Pure system [25]
XIIa 2.2 × 103 Pure system [26]
 A. Derbalah et al.

from porcine intestine or, less commonly, bovine intestine.

Chemically, UFH is an heterogeneous mixture of linear
polysaccharides that vary in chain length and sequence.
Chains are mainly constituted by alternating disaccharide
units of a uronic acid (l-iduronic acid or d-glucuronic acid)
and amino sugar (d-glucosamine) with variable degrees of
O- and N-sulfation and N-acetylation. Iduronic acid and glu-
cosamine disaccharide units are linked by a α-1,4 linkage to
the next residue, while glucuronic acid-containing units are
linked by β-1,4 linkage [46]. Each disaccharide unit of UFH
bears three or, less commonly, two sulfate groups. The dis-
tribution pattern of sulfate groups is heterogeneous among
UFH chains. The overall degree of sulfation of a UFH prepa-
ration also varies depending on the animal, organ source and
purification procedure [47].
The mixture of molecular weights (MW) of the UFH
chains is a key structural feature impacting its anticoagulant Fig. 1  MW distribution of pharmaceutical UFH from various tis-
activity (as described in Sects. 3.1.1 and 3.1.2). Pharmaceu- sue sources. a Bovine intestine heparin; b porcine intestine heparin.
tical UFH preparations consist of species with MW ranging Data were extracted from gel permeation chromatography plots in the
study by Tovar et al. [49] using WebPlotDigitzer [50]. MW was esti-
from 3 to 60 kDa, with an average MW between 11 and
mated by fitting a polynomial model to data of the standard curve.
20 kDa [48, 49]. The MW distributions of pharmaceutical Refractive index values were assumed to be linearly proportional
UFH preparations sourced from porcine or bovine intestine to the amount of heparin in the eluate. MW molecular weight, UFH
are shown in Fig. 1. unfractionated heparin
The presence of a specific pentasaccharide sequence with
a characteristic 3-O-sulfated glucosamine residue that can
bind AT with high affinity is another important structural
feature of UFH affecting anticoagulation activity. This AT K H:AT H:AT
kon
d
H + AT −
−−−
← −−−
− [H:AT]∗ −
→ ←−−−−
→− H:AT, (3)
binding sequence is present in approximately 30–50% of H:AT
koff
UFH molecules that are referred to as high-affinity UFH, in
contrast to low-affinity UFH species that do not exhibit this
K H:AT:Fa
sequence [51–54]. The MW distribution of high- and low- d kH:AT−Fa
(4)
H:AT + Fa −
−−−
← −−−−−
→
− H:AT:Fa ⟶ AT:Fa + H,
affinity UFH is essentially the same [53].
where KdH:AT represents the dissociation constant for the ini-
3.1 Pharmacodynamics of UFH tial weak complex formation equilibrium, kon H:AT
is the rate
constant for the conformational conversion of the initial
The major anticoagulant effect of UFH is mediated through weak complex to a stable complex, and kH:AT−Fa is the rate
binding and activation of the two key plasma serpins, namely constant of the reverse conformational change and represents
AT and HCII. Figure 2 provides a summary of the serpin- the stability of the final complex. The more stable conforma-
based mechanisms of the UFH anticoagulant effect. Additional tion of the complex ( H:AT ) is likely to be thermodynami-
non-serpin-mediated anticoagulant effects have been described cally favourable over the less stable conformation [H:AT]∗ ,
in the literature and are discussed in Sect. 3.1.3. which makes it unlikely to revert back to the unstable state.
Nevertheless, Olson et al. reports that a small koffH:AT
is pro-
3.1.1 UFH‑Mediated Activation of AT posed to explain the hyperbolic dependence of the observed
rate constant of the overall reaction on heparin concentration
Unfractionated heparin molecules containing the specific pen- [55]. The overall affinity (estimated as K­ D to represent both
tasaccharide sequence (high-affinity UFH) can bind to AT, components of the reaction) of UFH to AT is the combined
forming a complex that inactivates the activated clotting factor effects of the binding and rate constants. Finally, we see that
(Fa) more rapidly than the free AT (Eqs. 3 and 4). This binding UFH is conserved in these reactions. As shown in Table 2,
and activation of AT occurs in two steps (Eq. 3). Initially, a the synthetic pentasaccharide containing the minimum bind-
weak complex is formed in a rapid equilibrium, and a confor- ing sequence for AT has slightly less overall affinity than
mational change then converts the intermediate complex to a larger MW fragments containing the same binding sequence.
stable complex that is capable of rapid inhibition of clotting This suggests that the non-binding domain of the heparin
factors [55]. chain may have a role in stabilizing the final complex (lower
Pharmacology of Unfractionated Heparin

Fig. 2  UFH interaction with AT and HCII. Reactions are numbered tertiary pathway (slowest). Orange arrows represent reactions com-
according to their order of appearance in the text. Single-headed mon to both the primary and secondary pathways. UFH unfraction-
arrows represent irreversible reaction, while double-headed arrows ated heparin, AT antithrombin, HCII heparin cofactor II, IIa activated
represent reversible reactions. Light blue arrows represent reactions clotting factor II, Fa activated clotting factors IX, X, XI, XII, and
of the primary pathway (fastest), violet arrows represent reactions of VIIa:TF
the secondary pathway, and black arrows represent reactions of the

H:AT
koff values). UFH has an apparent lower overall affinity to a way that promotes recognition and interaction with the tar-
AT because the value presented in the table is the average for get activated clotting factor [60]. However, this mechanism
heparin chains containing and lacking AT binding sequence. insufficiently explains the full potential of UFH acceleration of
It has also been reported that as the MW increases, the pro- AT reactions, especially IIa inactivation, with the AT binding
portion of heparin chains having two binding sites for AT pentasaccharide alone minimally accelerating the rate of IIa
increases [56]. inactivation by AT (by less than twofold) [61]. A minimum of
Kinetic analysis of heparin-catalysed inactivation of clot- 18-saccharide units is required to demonstrate substantial IIa
ting factors by AT shows that the reaction occurs in two inactivation [61, 62], suggesting that a bridging mechanism is
steps (Eq. 4). The first step is reversible and involves the an additional catalyst of the IIa–AT interaction. The relative
formation of a ternary complex of UFH, AT, and the acti- contribution of the conformational change (which requires spe-
vated clotting factor. The second step is an irreversible rate- cific saccharide sequence) and bridging effects (dependent on
limiting dissociation of the ternary complex releasing free MW) is a function of the target protease. The bridging mecha-
UFH and inactive AT protease complexes [57, 58]. nism appears critical for catalysis of the inactivation of factors
The primary target proteases for the UFH:AT complex IIa, IXa and XIa, and hence inactivation of these proteases is
are IIa, Xa, and IXa. Additionally, to a lesser extent, the largely dependent on the MW of the heparin chain. In con-
complex can inactivate XIa and XIIa. Table 3 shows the trast, inactivation of factors Xa and XIIa can be significantly
second-order rate constants for the inhibition of activated catalysed by heparin fragments containing the pentasaccharide
clotting factors by the UFH:AT complex. AT binding sequence, although larger heparin fragments have
The primary mechanism by which UFH accelerates AT greater catalyst activity [63]. Table 4 compares the effects of
inhibition of proteases is through an allosterically induced MW of various high-affinity heparin fractions on acceleration
conformational change that exposes the active site of AT in of AT interaction with IIa and Xa. Figure 3 shows that small
heparin chains (up to 4.2 kDa) have much higher acceleratory
 A. Derbalah et al.

Table 2  Dissociation and rate constants for AT interaction with various heparin species
Heparin type MW (kDa) KDH:AT (M) KdH:AT (M) H:AT
kon ­(min−1) H:AT
koff ­(min−1) References

Synthetic pentasaccharide 1.8 3.6 × 10−8 2 × 10−7 24 × 103 54 [22]

High-affinity heparin fraction ~ 7.9a 9.7 × 10−9 2.9 × 10−7 31.2 × 103 12 [22]
UFH ~ 15 3.7 × 10−7 – – – [20]

AT antithrombin, MW molecular weight, UFH unfractionated heparin

a
 MW of a fraction of high-affinity heparin chains with reduced polydispersity

Table 3  Inhibition rate constants for UFH-catalysed AT clotting fac- K H:Fa

tor reactions H + Fa −
←
D
−−
−−→
− H:Fa, (5a)

Clotting factor kH:AT−Fa ­(M−1 ­min−1) References

H:Fa + AT ⇄ H:AT:Fa. (5b)
IIa 1.3 × 109 [20]
1.7 × 109 [24] 3.1.2 UFH‑Mediated Activation of HCII
Xa 2.4 × 108 [24]
IXa 3 × 108 [24] In contrast to AT, UFH binds to HCII non-specifically, with
XIa 3.6–9 × 105 [59] a minimum chain length of 13 saccharide units [41]. Both
XIIa 2.5 × 104 [26] high and low AT affinity heparins have similar HCII-medi-
ated IIa inactivation [71]. UFH binding induces a conforma-
UFH unfractionated heparin, AT antithrombin tional change in HCII identical to that produced in AT that
renders it a more rapid inactivator of IIa. HCII has minimal
contribution to the overall anticoagulant effect of UFH since
effect on Xa than IIa inhibition rate constants. It is important the affinity of UFH for AT is much higher. This explains why
to note that this figure does not represent the anti-Xa/anti-IIa heparin does not have substantial HCII-mediated anti-IIa
activity ratio, which is dependent on the rate of inhibition reac- activity, except under high concentrations [71, 72]. Interest-
tion (rather that the rate constants presented here). ingly, unlike AT, the UFH–HCII complex is able to inacti-
A secondary pathway of UFH catalysis of clotting factor vate fibrin-bound thrombin, possibly because the bridging
inactivation involves binding of UFH to the activated clotting of HCII and IIa on the same heparin chain is not required
factor. This complex interacts with AT to form a ternary com- [73] and that HCII binds to IIa distinct from its complex
plex (Eqs. 5a and 5b) that dissociates to release free heparin with fibrin [74].
and inactive the AT clotting factor complex [67, 68]. Bind- Binding of heparin to HCII involves rapid formation of an
ing of UFH to activated clotting factors may only take place unstable encounter complex that undergoes a conformational
in AT-saturating concentration of UFH because binding AT change, yielding a more stable activated complex (Eq. 6).
is the kinetically preferred pathway, as illustrated in Table 5, The overall affinity of UFH to HCII ( KDH:HCII ) is the com-
where the affinity of heparin to AT and different proteases are posite of the combined effect of the dissociation constant of
compared. the encounter complex KdH:HCII and the rate constants for the

Table 4  Second-order rate MW of heparin frac- kH:AT−Xa ­(M−1 References kH:AT−IIa ­(M−1 ­min−1) References
constants of clotting factor tion (kDa) ­min−1)
inactivation by AT catalysed
by various high-affinity heparin 1.7 1.2 × 107 [64] 1.46 × 104 [22]
fractions
2.7–3.2 1.7 × 107 3 × 105 [65]
4.2–4.3 3.4 × 107 2 × 107
6–6.2 7.7 × 107 0.24 × 109 [66]
9 2.6 × 108 0.98 × 109
16.5 3 × 108 1.39 × 109
22.5 4.2 × 108 1.91 × 109
32 4.2 × 108 1.76 × 109

AT antithrombin, MW molecular weight

Pharmacology of Unfractionated Heparin

Fig. 3  Molecular weight
dependence of the catalytic
activity of heparin on AT-
mediated inactivation of clotting
factors Xa and IIa. Blue and
pink lines (presented on the
right y-axis) represent the
acceleration of inhibition of IIa
and Xa, respectively. Values
are calculated as the ratios of
the second-order rate constants
of catalysed to uncatalaysed
reactions performed by the
same investigator under the
same experimental conditions
(references see Table 4). The
green line (presented on the left
y-axis) represents the ratio of
folds of Xa to IIa acceleration of
second-order rate constants

Table 5  Dissociation Agent KD (M) References HCII:IIa

kon
constants of heparin:AT H:HCII + IIa ⇄ H:HCII:IIa ⟶ HCII:IIa + H. (7)
and heparin:clotting factor AT 5.74 × 10−8 [69, 70]
complexes
IIa 8 × 10−7
Xa 8.73 × 10−6 3.1.3 Non‑Serpin‑Dependent Mechanism
IXa 2.58 × 10−7
XIa – – The in  vitro anticoagulant effect of UFH can be com-
XIIa – – pletely neutralised by polybrene (a polymerised quater-
nary ammonium salt). However, ex  vivo, UFH has an
Values reported are for high- anticoagulant effect that cannot be abolished completely
affinity heparin fractions with
an average molecular weight
by polybrene. This unneutralisable effect is mediated by
(6.5 kDa) stimulation of the release and potentiation of the activ-
AT antithrombin ity of a potent endogenous anticoagulant, namely tissue
factor pathway inhibitor (TFPI). When measured using
a diluted prothrombin time assay, the TFPI-dependent
forward and backward conformational changes, kon H:HCII
and mechanism contributes to approximately one-third of the
H:HCII
koff respectively. Table 6 shows chain length dependence anticoagulant effects of UFH [77]. The stimulation of TFPI
of HCII affinity to heparin. release is dose-dependent, with an approximately 3.7 ng/
K H:HCII H:HCII
kon
mL increase in TFPI concentration per 100 IU of UFH at
a dose level of 5000 IU. However, repeated and continued
d
−−−−
H + HCII ← −−−−→ −−−−−−−
−− [H:HCII]∗ ← →− H:HCII. (6)
intravenous administration of UFH results in an attenua-
H:HCII
koff

tion of the releases of TFPI, and subsequently a decrease

HCII, when activated by UFH, is 1000- to 3000-fold in its contribution to UFH prolongation of diluted pro-
more active than the native protein. Inhibition of IIa occurs thrombin time [78].
in two steps. The first step involves formation of a ternary TFPI inhibits factor Xa as well as VIIa:TF complexes
UFH:HCII:IIa complex, which then irreversibly dissociates, in a reversible two-step reaction mechanism, as shown in
releasing free UFH and inactive HCII:IIa complex (Eq. 7). Eqs. 8 and 9 [79]. However, it is not clear if the quaternary
The irreversible dissociation is the rate-limiting step of the complex [TFPI:XA] ∶ [VIIa:TF] has a different fate other
inhibition reaction and has a second-order rate constant than dissociating back to the two forming subcomplexes.
( kon
HCII:IIa
 ) of 1.4 × 107–1.14 × 109 M−1 min−1 [41, 66, 75, 76]. The kinetics of TFPI inhibition on Xa and VIIa:TF under
 A. Derbalah et al.

Table 6  Dissociation and rate constants for HCII interaction with var- Table 7  Effect of UFH on the kinetics of TFPI-mediated clotting fac-
ious heparin species [41] tor inhibition
Heparin type Chain length (saccha- KdH:HCII (M) Reaction kon ­(M−1 min−1) koff ­(min−1) References
ride units)
8 0.86 × 109 0.04 [80]
Fractionated heparin 14 27.5 × 10−4 8 + 0.05 U/mL UFH 2.1 × 109 0.04
16 12 × 10−5 9 5.8 × 108 –
18 11 × 10−5 9 + 0.05 U/mL UFH 1.6 × 108 –
20 59 × 10−6
26 45 × 10−6 UFH unfractionated heparin, TFPI tissue factor pathway inhibitor
UFH ~ 50 26 × 10−6

UFH unfractionated heparin, HCII heparin cofactor II disease state, and concomitant administration of blood and/
or blood products) [85].
Protamine titration is often referred to as the gold stand-
near-physiologic conditions in the presence and absence ard for the measurement of UFH concentrations [86]. How-
of UFH are summarised in Table 7. ever, protamine sulfate binding to UFH is dependent on
TFPI:Xa
the degree of sulfation of chains [87] and also on MW, i.e.
kon
TFPI + Xa −
−−−−−−−→ binding only to chains above 4.2 kDa, with a proportional
← −− TFPI:Xa, (8)
TFPI:Xa
koff increase in affinity as MW increases [88]. As a result, the
amount of protamine required to neutralise the anti-IIa activ-
[TFPI:Xa]∶[VIIa:TF]
ity of UFH is less than that required to neutralise its anti-Xa
activity. UFH-induced prolongation of plasma clotting time,
kon
−−−−−−−−−−−−−−−−−−−
TFPI:Xa + VIIa:TF ← →
− [TFPI:XA] ∶ [VIIa:TF].
[TFPI:Xa]∶[VIIa:TF]
koff as measured by aPTT or thromboelastography, also requires
(9) more protamine in order to be normalised than that required
3.2 Laboratory Monitoring to neutralise the anti-IIa activity of UFH [51]. This means
that protamine titration should not be considered a reliable
Laboratory monitoring is recommended to guide UFH measure of UFH concentration.
dose adjustment to achieve optimal therapeutic outcomes Assays of the anti-Xa and anti-IIa activities are being
[81]. An ideal laboratory monitoring test should capture increasingly used in clinical practice as an alternative to pro-
the full spectrum of UFH activity discussed above, have tamine titration that can be easily automated and standard-
precise and well-standardised reference ranges across ised. However, both tests are only sensitive to the fractions
laboratories and reagents, be inexpensive and easily per- of UFH that possess the AT binding sequence. They are also
formed, in addition to having a well-characterised relation- sensitive to AT concentration (methods that do not involve
ship to clinical outcomes (including adverse effects such as the addition of exogenous AT), making them a less-specific
bleeding). Unfortunately, such a test does not exist, mak- measure of UFH effect [89]. As with protamine titration,
ing laboratory monitoring of UFH problematic. anti-Xa and anti-IIa assays are more accurately described
Activated partial thromboplastin time (aPTT) is the as a measure of UFH PD effect rather than concentration.
most widely used measure of coagulation system response Various assays have recently been developed for quan-
to UFH. Its use is supported by its ease and availability of tification of heparin concentration in matrices, including
measurement. However, aPTT is a non-specific measure of plasma directly and independently of its anticoagulant
blood coagulability and hence is not specific to UFH. It is activity [90–93]; however, none of these assays has been
affected by concentrations of clotting factors of the intrinsic evaluated for ex vivo measurement of UFH and their clinical
and common coagulation pathways. Thus, the test shows a utility remains unclear.
baseline variability corresponding to the variability inherent
to the haemostatic system itself [3]. Additionally, the test
shows a significant sensitivity to many preanalytic (sampling 3.3 PK of UFH
time, storage, citrate concentration, centrifugation condi-
tions) and analytic variables (thromboplastin reagent, clot The mode of UFH clearance from plasma is poorly under-
detector) [82]. It is not surprising that only 1–12% of vari- stood, mainly because of the difficulty in characterising the
ability in aPTT can be explained by UFH dose [83, 84]. The metabolic and excretory fate of a heterogeneous polymer.
remaining variability in such a response measure is likely to UFH binds to and is internalised by cells of the reticuloen-
be related, at least in part, to the PK, PD, and coagulation dothelial system, especially in the liver, spleen and kidney,
system variability (which is affected by factors such as age, undergoing desulfation and depolymerisation. Binding is
Pharmacology of Unfractionated Heparin

saturable, reversible and dependent on the MW of UFH, Heparin binds to several plasma proteins that modulate
but not AT affinity (the binding equilibrium constant KD is its activity and are responsible for much of the variability
0.23 × 10−6 M for UFH and 4.3 × 10−6 M for low molecular in anticoagulant effect. One of the most important of these
weight heparin of MWs from 1.5 to 8 kDa) [94]. UFH is proteins is platelet factor 4 (PF4). PF4 binds non-specifically
recoverable from urine in a partially desulfated and depoly- to UFH chains but with high affinity (KD 3 × 10−8 M) [100].
merised form that retains some anticoagulant activity, yet Such high affinity means that UFH will bind AT or HCII
very few intact chains are excreted in urine [95]. only when it is abundant enough to saturate PF4. Therefore,
The inherent mixture properties of UFH complicate our the anticoagulant activity of UFH will be largely depend-
understanding of the kinetics of its activity. For example, the ent on the level of PF4 and other UFH-binding proteins.
decline in UFH activity (as measured by aPTT or IIa clotting Low-affinity heparin added to ex vivo samples from patients
time assays) after an intravenous bolus injection appears to receiving UFH results in an increase in its anti-Xa activity,
be exponential with time. However, the decline is less steep indicating that protein binding is potentially an important
at higher UFH doses regardless of the assay method used effector of the anticoagulant effect of UFH [101]. Other
[96]. Additionally, the disappearance of UFH activity (as plasma proteins that can also bind and modulate the antico-
measured by aPTT) after discontinuation of a continuous agulant effect of UFH are histidine-rich glycoprotein (KD
intravenous infusion is less steep compared with what would 7 × 10−10 M) [102] and vitronectin (KD 4 × 10−8 M).
be predicted from pre-infusion bolus data [97].
The mechanisms underpinning this complex dose-, time-, 3.4 Discussion
and assay-dependent behaviour are unknown, although sev-
eral models have been proposed to explain these actions. Unfractionated heparin is one of the most commonly used
The dual elimination mechanism model suggests that UFH is anticoagulant drugs. Its short duration of action and revers-
cleared via a combination of first-order and saturable-order ibility of its anticoagulant effect by protamine sulfate have
mechanisms, which may result in the elimination of different supported its widespread adoption as the anticoagulant of
MW species at different rates, perhaps with higher MW spe- choice in surgical operations. UFH is also commonly used
cies being more dependent on a saturable pathway [98]. This for the prevention and treatment of venous thromboembolic
model possibly explains the relatively low biological bio- disorders. However, a major limitation associated with its
availability observed when UFH is administered subcutane- use is the unpredictable dose–response relationship. As a
ously at low doses compared with high doses (as determined result, accurate determination of optimal doses is difficult.
from the area under the curve [AUC] of anti-Xa activity The information presented in this review provides a compre-
versus time). The model suggests that the dose-normalised hensive analysis of potential mechanisms underpinning the
maximum plasma concentrations achieved after low doses observed high variability in response.
are lower than those of higher doses, resulting in higher The heterogeneous structure of UFH is a primary rea-
clearance mediated by the saturable clearance pathway, and son for its complicated PK profile. Different MW species
hence lesser AUC of plasma anti-Xa versus time [98]. are cleared from plasma at different rates and by different
Another model that describes the kinetic behaviour mechanisms, meaning that the relative abundance of each of
of UFH is the metabolite-inhibited elimination model. these species changes continuously over time and after each
This model suggests that heparin is metabolised to inac- dose. Measurement of the MW profile of in vivo UFH over
tive metabolites that compete with it for the metabolising time appears impractical. Alternatively, some researchers
enzymes [99]. This model describes the time-dependent have used methods to quantify the total amount of all UFH
nature of the decline in biological activity and explains the species, using radiolabelled UFH for instance. This prac-
longer infusion time required to reach steady state (as meas- tice is suboptimal because it does not delineate the change
ured by IIa clotting time) compared with what would be pre- in amounts of the various MW components of UFH and is
dicted from the apparent (anti-IIa activity-based) half-life. also impractical in practice. Monitoring UFH through meas-
It has also been suggested that heparin may be metabolised ures of anticoagulant activity (anti-Xa, anti-IIa) in plasma
to an active, more slowly eliminated metabolite that accu- is even less informative, not only because it sums up the
mulates over time [96]. All of these models describe the total effects of all species but also because it captures the
observed behaviour of UFH, but they are all experimentally variability within the haemostatic system itself, adding more
unprovable unless a method to quantify UFH concentration noise to the already indefinite signal. Such assays also do
in plasma exists. Note that this model does not imply that the not capture the effect produced by low-affinity UFH chains.
dose-normalised maximum plasma concentrations achieved Protamine sulfate has been used to reverse the effect of
after low doses are lower than those of higher doses, which UFH in cases of toxicity, or when the effect is no longer
provides a potential point for model comparison. required, as well as to quantify its amount in plasma. Prota-
mine has a higher affinity for longer UFH chains than shorter


Fig. 4  Interaction of UFH with the haemostatic system. Dashed lines represent a stimulation of release reaction, and solid lines represent mass transfer reaction. Colours of the arrows indicate
the relative speed of reactions: blue represents the fastest (primary pathway), green represents the secondary pathway, orange represents reactions common to the primary and secondary path-
ways, and black represents the tertiary pathway. Blue squares represent endogenous species, yellow squares represent terminal states, and orange squares represent intermediary complexes. UFH
unfractionated heparin, AT antithrombin, TF tissue factor, TFPI tissue factor pathway inhibitor, HCII heparin cofactor II, IIa, VIIa, IXa, Xa, XIa, XIIa: activated clotting factors
A. Derbalah et al.
Pharmacology of Unfractionated Heparin

UFH chains. Therefore, it more rapidly neutralises the anti- comprehensive quantitative approach to describing the PK
IIa effects of UFH than the anti-Xa effects. It will also not and PD of UFH that could serve as a basis for a proposed
alter the anticoagulant effect of UFH produced by stimula- model. A schematic presentation of a comprehensive PD
tion of TFPI release. The amount of protamine required to model for the anticoagulant effect of UFH is shown in
neutralise UFH in plasma will vary by the method used to Fig. 4, which shows the complete set of biochemical reac-
detect the endpoint (IIa clotting time vs. Xa clotting time tions known to be involved in UFH interactions with the
vs. aPTT). Additionally, protamine titration also captures haemostatic system. The kinetics of these reactions have
the noise of the inherent variability of the haemostatic sys- been comprehensively reviewed and described in this
tem. Therefore, estimates of kinetic parameters vary by the article. Such information is an essential prerequisite for a
method used to detect the endpoint. quantitative systems model that aims to describe the phar-
The PK profile of UFH is further complicated by mul- macology of UFH. For the proposed model, Fig. 4 serves
tifaceted PD behaviour. The main mechanism of the UFH as framework to construct a system of ordinary differential
anticoagulant effect can be explained through an enzyme- equations (ODEs), and the kinetic constants outlined here
substrate model. In this model, UFH functions as an enzyme represent reasonable initial conditions of model states.
that accelerates the reaction of two substrates (AT and IIa, With such a model, it would be feasible to closely
for instance). The enzyme (UFH) is released after each monitor the series of reactions that occur when the sys-
reaction is available, to catalyse more cycles of reactions. tem is perturbed by UFH, leading to a PD response. A
However, such a model is complicated by the fact that dif- lot of information can be learnt from this investigation.
ferent MW species act differently. For instance, high AT For example, it would be possible to identify which of
affinity UFH species only catalyse AT reactions, low AT the competing PK models of UFH (see Sect. 3.3) more
affinity species bind HCII, especially the high MW chains, accurately describe the actual PK behaviour of the UFH
and stimulate the release and activity of TFPI. In addition, mixture. The model can also be used to investigate the
these fragments also potentiate high-affinity activity by com- net effect of UFH on TFPI activity and what conditions
petition for binding to PF4 and other heparin-binding plasma (if any) can change this effect. Such information can then
proteins. be used to build a more informed dose-response model of
The PD of UFH are further complicated by the fact that UFH that, when validated, can be used to optimise UFH
some of its reactions might counteract the effect of each dosing. Dose optimisation in this context refers to predic-
other. For example, UFH heparin binds to TFPI and stimu- tion of the dose that produces an optimal therapeutic effect
lates its binding to, and inactivation of, Xa as a required (i.e. anticoagulation response) and minimises dose-related
pre-emptive step to inactivation of the TF:VIIa complex. adverse effects (i.e. bleeding or treatment failure). Of note,
At the same time, UFH may reduce TFPI activity through optimised dosing is unlikely to prevent other adverse
a reduction in the amount of Xa available to TFPI (by effects that do not have a well-defined relationship with
enhancing AT-mediated Xa inactivation). The net effect the UFH dose, such as heparin-induced thrombocytopenia.
is not known (potentiation or inhibition) and may vary
over time and in relation to heparin concentration. It is Compliance with Ethical Standards 
possible that factors such as Xa and AT concentrations,
for instance, will be determinants of net effect. The impact Funding  No external funding was used in the preparation of this manu-
script.
of such opposing pathways on the overall in vivo antico-
agulant effect of UFH is also unknown. Conditions that
Conflict of Interest Abdallah Derbalah, Stephen Duffull, Fiona Ne-
favour one of these effects over the other, if any, are likely wall, Katie Moynihan, and Hesham Al-Sallami declare that they have
to contribute to the observed unexplained variability in no potential conflicts of interest that might be relevant to the contents
UFH response of this manuscript.

3.5 Conclusions and Future Perspectives References

The inherent mixture properties of UFH complicate our 1. Jacqmin P, Snoeck E, van Schaick EA, Gieschke R, Pillai P,
Steimer J-L, et al. Modelling response time profiles in the
understanding of its PK and PD, and hence the ability to absence of drug concentrations: definition and performance
identify an optimal dose. These UFH properties may lend evaluation of the K-PD model. J Pharmacokinet Pharmacodyn.
it to exploration using a systems pharmacology approach 2007;34(1):57–85.
that may facilitate understanding about behaviours 2. Gabrielsson J, Jusko WJ, Alari L. Modeling of dose-response-
time data: four examples of estimating the turnover parameters
that are not yet understood. This review has provided a
 A. Derbalah et al.

and generating kinetic functions from response profiles. Biop- 23. Jesty J. Measurement of the kinetics of inhibition of activated
harm Drug Dispos. 2000;21(2):41–52. coagulation factor X in human plasma: The effect of plasma and
3. Duffull SB. Is the ideal anticoagulant a myth? Expert Rev Clin inhibitor concentration. Anal Biochem. 1986;152(2):402–11.
Pharmacol. 2012;5(3):231–6. 24. Jordan RE, Oosta GM, Gardner WT, Rosenberg RD. The
4. Al-Sallami H, Newall F, Monagle P, Ignjatovic V, Cranswick kinetics of hemostatic enzyme-antithrombin interactions in
N, Duffull S. Development of a population pharmacokinetic– the presence of low molecular weight heparin. J Biol Chem.
pharmacodynamic model of a single bolus dose of unfrac- 1980;255(21):10081–90.
tionated heparin in paediatric patients. Br J Clin Pharmacol. 25. Scott CF, Schapira M, James HL, Cohen AB, Colman RW. Inac-
2016;82(1):178–84. tivation of factor XIa by plasma protease inhibitors: predominant
5. Jia Z, Tian G, Ren Y, Sun Z, Lu W, Hou X. Pharmacokinetic role of alpha 1-protease inhibitor and protective effect of high
model of unfractionated heparin during and after cardiopulmo- molecular weight kininogen. J Clin Investig. 1982;69(4):844–52.
nary bypass in cardiac surgery. J Transl Med. 2015;13(1):45. 26. Pixley R, Schapira M, Colman R. Effect of heparin on the inac-
6. Delavenne X, Ollier E, Chollet S, Sandri F, Lanoiselée J, Hodin tivation rate of human activated factor XII by antithrombin III.
S, et al. Pharmacokinetic/pharmacodynamic model for unfrac- Blood. 1985;66(1):198–203.
tionated heparin dosing during cardiopulmonary bypass. Br J 27. Conard J, Brosstad F, Lie Larsen M, Samama M, Abildgaard
Anaesth. 2017;118(5):705–12. U. Molar antithrombin concentration in normal human plasma.
7. Brunet P, Simon N, Opris A, Faure V, Lorec-Penet AM, Por- Haemostasis. 1983;13(6):363–8.
tugal H, et al. Pharmacodynamics of unfractionated heparin 28. Collen D, Schetz J, de Cock F, Holmer E, Verstraete M. Metabo-
during and after a hemodialysis session. Am J Kidney Dis. lism of antithrombin III (heparin cofactor) in man: effects of
2008;51(5):789–95. venous thrombosis and of heparin administration. Eur J Clin
8. Bonate PL. The art of modeling. Pharmacokinetic–pharmaco- Investig. 1977;7(1):27–35.
dynamic modeling and simulation. Boston: Springer US; 2011. 29. Andrew M, Paes B, Milner R, Johnston M, Mitchell L, Tollefsen
p. 1–60. DM, et al. Development of the human coagulation system in the
9. Wajima T, Isbister GK, Duffull SB. A comprehensive model for full-term infant. Blood. 1987;70(1):165–72.
the humoral coagulation network in humans. Clin Pharmacol 30. Lu W, Mant T, Levy JH, Bailey JM. Pharmacokinetics of recom-
Ther. 2009;86(3):290–8. binant transgenic antithrombin in volunteers. Anesth Analg.
10. Hoffman MM, Monroe DM. Rethinking the coagulation cas- 2000;90(3):531–4.
cade. Curr Hematol Rep. 2005;4(5):391–6. 31. Moffett BS, Diaz R, Galati M, Mahoney D, Teruya J, Yee DL.
11. Rosenberg RD. Role of heparin and heparinlike molecules in Population pharmacokinetics of human antithrombin concentrate
thrombosis and atherosclerosis. Fed Proc. 1985;44(2):404–9. in paediatric patients. Br J Clin Pharmacol. 2017;83(11):2450–7.
12. Morawitz P. Die Chemie der Blutgerinnung. Ergebnisse der 32. DeJongh J, Frieling J, Lowry S, Drenth H-J. Pharmacokinetics of
Physiologie. 1905;4(1):307–422. recombinant human antithrombin in delivery and surgery patients
13. Brinkhous KM, Smith HP, Warner ED, Seegers WH. The with hereditary antithrombin deficiency. Clin Appl Thromb
inhibition of blood clotting: an unidentified substance which Hemost. 2013;20(4):355–64.
acts in conjunction with heparin to prevent the conversion of 33. Lam LSL, Regoeczi E, Hatton MWC. In  vivo behaviour of
prothrombin into thrombin. Am J Physiol Legacy Content. some antithrombin III–protease complexes. Br J Exp Pathol.
1939;125(4):683–7. 1979;60(2):151–60.
14. Gerendas M. Inactivation of thrombin. Nature. 1946;157:837. 34. Esposito RA, Culliford AT, Colvin SB, Thomas SJ, Lackner H,
15. Gerendas M. Inactivation and stabilization of thrombin. Hung Spencer FC. Heparin resistance during cardiopulmonary bypass.
Acta Physiol. 1948;1(4–5):97–115. The role of heparin pretreatment. J Thorac Cardiovasc Surg.
16. Owen WG. Evidence for the formation of an ester between 1983;85(3):346–53.
thrombin and heparin cofactor. Biochim Biophys Acta (BBA) 35. Porter P, Porter MC, Shanberge JN. Heparin cofactor and plasma
Protein Struct. 1975;405(2):380–7. antithrombin in relation to the mechanism of inactivation of
17. Carlson TH. Clearance of thrombin in vivo: significance of thrombin by heparin. Clin Chim Acta. 1967;17(2):189–200.
alternative pathways. Mol Cell Biochem. 1986;71(2):97–105. 36. Sie P, Dupouy D, Pichon J, Boneu B. Constitutional heparin co-
18. Bock SC. Antithrombin and the Serpin Family. In: Marder VJA, factor II deficiency associated with recurrent thrombosis. Lancet.
Bennett WC, Schulman JS, White S, Gilbert C, editors. Hemosta- 1985;2(8452):414–6.
sis and thrombosis: basic principles and clinical practice. 6th ed. 37. Tran TH, Duckert F. Heparin cofactor II determination–levels
Philadelphia: Lippincott Williams and Wilkins; 2013. p. 286–99. in normals and patients with hereditary antithrombin III defi-
19. Jesty J. The kinetics of inhibition of thrombin by antithrombin ciency and disseminated intravascular coagulation. Thromb
in the presence of components of the hemostatic system. Blood. Haemost. 1984;52(2):112–6.
1985;66(5):1189–95. 38. Baglin TP, Carrell RW, Church FC, Esmon CT, Huntington JA.
20. Maaroufi RM, Jozefowicz M, Tapon-Bretaudière J, Fischer Crystal structures of native and thrombin-complexed heparin
A-M. Mechanism of thrombin inhibition by antithrombin and cofactor II reveal a multistep allosteric mechanism. Proc Natl
heparin cofactor II in the presence of heparin. Biomaterials. Acad Sci. 2002;99(17):11079–84.
1997;18(3):203–11. 39. Tovar AMF, de Mattos DA, Stelling MP, Sarcinelli-Luz BSL,
21. Downing MR, Bloom JW, Mann KG. Comparison of the inhibi- Nazareth RA, Mourão PAS. Dermatan sulfate is the pre-
tion of thrombin by three plasma protease inhibitors. Biochem- dominant antithrombotic glycosaminoglycan in vessel walls:
istry. 1978;17(13):2649–53. Implications for a possible physiological function of heparin
22. Olson ST, Bjork I, Sheffer R, Craig PA, Shore JD, Choay J. Role cofactor II. Biochim Biophys Acta (BBA) Mol Basis Dis.
of the antithrombin-binding pentasaccharide in heparin accel- 2005;1740(1):45–53.
eration of antithrombin-proteinase reactions. Resolution of the 40. Vinazzer H. Heparin cofactor II: structure, function, and clini-
antithrombin conformational change contribution to heparin rate cal importance. In: Sas G, editor. The biology of antithrombins.
enhancement. J Biol Chem. 1992;267(18):12528–38. Boca Raton: CRC Press; 1990. p. 141–55.
Pharmacology of Unfractionated Heparin

41. O’Keeffe D, Olson ST, Gasiunas N, Gallagher J, Baglin TP, 61. Olson ST, Bjork I. Predominant contribution of surface
Huntington JA. The heparin binding properties of heparin approximation to the mechanism of heparin acceleration of the
cofactor II suggest an antithrombin-like activation mechanism. antithrombin–thrombin reaction. Elucidation from salt concen-
J Biol Chem. 2004;279(48):50267–73. tration effects. J Biol Chem. 1991;266(10):6353–64.
42. Sie P, Dupouy D, Pichon J, Boneu B. Turnover study of 62. Holmer E, Lindahl U, Bäckström G, Thunberg L, Sandberg H,
heparin cofactor II in healthy man. Thromb Haemost. Söderström G, et al. Anticoagulant activities and effects on plate-
1985;54(3):635–8. lets of a heparin fragment with high affinity for antithrombin.
43. McLean J. The discover y of heparin. Circulation. Thromb Res. 1980;18(6):861–9.
1959;19(1):75–8. 63. Holmer E, Kurachi K, Soderstrom G. The molecular-weight
44. Oduah EI, Linhardt RJ, Sharfstein ST. Heparin: past, present, and dependence of the rate-enhancing effect of heparin on the inhibi-
future. Pharmaceuticals (Basel). 2016;9(3):38. tion of thrombin, factor Xa, factor IXa, factor XIa, factor XIIa and
45. Nader HB, Chavante SF, dos-Santos EA, Oliveira TW, de-Paiva kallikrein by antithrombin. Biochem J. 1981;193(2):395–400.
JF, Jeronimo SM, et al. Heparan sulfates and heparins: similar 64. Ellis V, Scully MF, Kakkar VV. The relative molecular mass
compounds performing the same functions in vertebrates and dependence of the anti-factor Xa properties of heparin. Biochem
invertebrates? Braz J Med Biol Res. 1999;32(5):529–38. J. 1986;238(2):329–33.
46. Rodén L, Ananth S, Campbell P, Curenton T, Ekborg G, Man- 65. Hoylaerts M, Owen WG, Collen D. Involvement of heparin
zella S, et al. Heparin—an introduction. In: Lane DA, Björk I, chain length in the heparin-catalyzed inhibition of thrombin by
Lindahl U, editors. Heparin and related polysaccharides. Boston: antithrombin III. J Biol Chem. 1984;259(9):5670–7.
Springer US; 1992. p. 1–20. 66. Scully MF, Ellis V, Kakkar VV. Comparison of the molecular
47. Bianchini P, Liverani L, Mascellani G, Parma B. Heterogene- mass dependency of heparin stimulation of heparin cofactor
ity of unfractionated heparins studied in connection with spe- II:thrombin interaction to antithrombin III:thrombin interaction.
cies, source, and production processes. Semin Thromb Hemost. Thromb Res. 1987;46(3):491–502.
1997;23(1):3–10. 67. Griffith MJ. Kinetics of the heparin-enhanced antithrombin III/
48. Mulloy B, Hogwood J, Gray E. Assays and reference materi- thrombin reaction. Evidence for a template model for the mecha-
als for current and future applications of heparins. Biologicals. nism of action of heparin. J Biol Chem. 1982;257(13):7360–5.
2010;38(4):459–66. 68. Machovich R. Mechanism of action of heparin through thrombin
49. Tovar AM, Santos GR, Capille NV, Piquet AA, Glauser BF, on blood coagulation. Biochim Biophys Acta. 1975;412(1):13–7.
Pereira MS, et  al. Structural and haemostatic features of 69. Jordan RE, Oosta GM, Gardner WT, Rosenberg RD. The binding
pharmaceutical heparins from different animal sources: chal- of low molecular weight heparin to hemostatic enzymes. J Biol
lenges to define thresholds separating distinct drugs. Sci Rep. Chem. 1980;255(21):10073–80.
2016;6:35619. 70. Jordan RE, Oosta GM, Gardner WT, Rosenberg RD. The
50. Rohatgi A. WebPlotDigitizer. 4.1 ed., Austin; 2018. kinetics of hemostatic enzyme-antithrombin interactions in
51. Hogwood J, Mulloy B, Gray E. Precipitation and neutralization the presence of low molecular weight heparin. J Biol Chem.
of heparin from different sources by protamine sulfate. Pharma- 1980;255(21):10081–90.
ceuticals (Basel). 2017;10(3):E59. 71. Sie P, Petitou M, Lormeau JC, Dupouy D, Boneu B, Choay J.
52. Lam LH, Silbert JE, Rosenberg RD. The separation of active Studies on the structural requirements of heparin for the catalysis
and inactive forms of heparin. Biochem Biophys Res Commun. of thrombin inhibition by heparin cofactor II. Biochim Biophys
1976;69(2):570–7. Acta. 1988;966(2):188–95.
53. Hook M, Bjork I, Hopwood J, Lindahl U. Anticoagulant activity 72. Petitou M, Lormeau JC, Perly B, Berthault P, Bossennec V, Sie
of heparin: separation of high-activity and low-activity heparin P, et al. Is there a unique sequence in heparin for interaction with
species by affinity chromatography on immobilized antithrombin. heparin cofactor II? Structural and biological studies of heparin-
FEBS Lett. 1976;66(1):90–3. derived oligosaccharides. J Biol Chem. 1988;263(18):8685–90.
54. Andersson LO, Barrowcliffe TW, Holmer E, Johnson EA, Sims 73. Sheehan JP, Tollefsen DM, Sadler JE. Heparin cofactor II is regu-
GE. Anticoagulant properties of heparin fractionated by affinity lated allosterically and not primarily by template effects. Studies
chromatography on matrix-bound antithrombin iii and by gel with mutant thrombins and glycosaminoglycans. J Biol Chem.
filtration. Thromb Res. 1976;9(6):575–83. 1994;269(52):32747–51.
55. Olson ST, Srinivasan KR, Bjork I, Shore JD. Binding of high 74. Baglin TP, Carrell RW, Church FC, Esmon CT, Huntington JA.
affinity heparin to antithrombin III. Stopped flow kinetic studies Crystal structures of native and thrombin-complexed heparin
of the binding interaction. J Biol Chem. 1981;256(21):11073–9. cofactor II reveal a multistep allosteric mechanism. Proc Natl
56. Rosenberg RD, Jordan RE, Favreau LV, Lam LH. Highly active Acad Sci USA. 2002;99(17):11079–84.
heparin species with multiple binding sites for antithrombin. Bio- 75. Tollefsen DM, Majerus DW, Blank MK. Heparin cofactor II.
chem Biophys Res Commun. 1979;86(4):1319–24. Purification and properties of a heparin-dependent inhibitor of
57. Olson ST, Shore JD. Demonstration of a two-step reaction thrombin in human plasma. J Biol Chem. 1982;257(5):2162–9.
mechanism for inhibition of alpha-thrombin by antithrombin III 76. Monagle P, Berry L, O’Brodovich H, Andrew M, Chan A. Cova-
and identification of the step affected by heparin. J Biol Chem. lent heparin cofactor II-heparin and heparin cofactor II-dermatan
1982;257(24):14891–5. sulfate complexes. Characterization of novel anticoagulants. J
58. Craig PA, Olson ST, Shore JD. Transient kinetics of heparin-cat- Biol Chem. 1998;273(50):33566–71.
alyzed protease inactivation by antithrombin III. Characterization 77. Abildgaard U, Lindahl AK, Sandset PM. Heparin requires both
of assembly, product formation, and heparin dissociation steps in antithrombin and extrinsic pathway inhibitor for its anticoagulant
the factor Xa reaction. J Biol Chem. 1989;264(10):5452–61. effect in human blood. Haemostasis. 1991;21(4):254–7.
59. Scott C, Colman R. Factors influencing the acceleration of 78. Hansen JB, Sandset PM, Huseby KR, Huseby NE, Nordoy A.
human factor XIa inactivation by antithrombin III. Blood. Depletion of intravascular pools of tissue factor pathway inhibi-
1989;73(7):1873–9. tor (TFPI) during repeated or continuous intravenous infusion of
60. Jin L, Abrahams JP, Skinner R, Petitou M, Pike RN, Carrell RW. heparin in man. Thromb Haemost. 1996;76(5):703–9.
The anticoagulant activation of antithrombin by heparin. Proc
Natl Acad Sci USA. 1997;94(26):14683–8.
 A. Derbalah et al.

79. Broze GJ Jr, Warren LA, Novotny WF, Higuchi DA, Girard JJ, that operates in highly competitive media. J Am Chem Soc.
Miletich JP. The lipoprotein-associated coagulation inhibitor 2013;135(8):2911–4.
that inhibits the factor VII-tissue factor complex also inhibits 91. Warttinger U, Giese C, Harenberg J, Holmer E, Kramer R. A fluo-
factor Xa: insight into its possible mechanism of action. Blood. rescent probe assay (Heparin Red) for direct detection of heparins
1988;71(2):335–43. in human plasma. Anal Bioanal Chem. 2016;408(28):8241–51.
80. Jesty J, Wun TC, Lorenz A. Kinetics of the inhibition of factor 92. Li G, Yang B, Li L, Zhang F, Xue C, Linhardt RJ. Analysis of
Xa and the tissue factor-factor VIIa complex by the tissue factor 3-O-sulfo group-containing heparin tetrasaccharides in heparin
pathway inhibitor in the presence and absence of heparin. Bio- by liquid chromatography-mass spectrometry. Anal Biochem.
chemistry. 1994;33(42):12686–94. 2014;455:3–9.
81. Hirsh J, Raschke R. Heparin and low-molecular-weight heparin: 93. Yoshimi Y, Yagisawa Y, Yamaguchi R, Seki M. Blood heparin
the Seventh ACCP Conference on Antithrombotic and Throm- sensor made from a paste electrode of graphite particles grafted
bolytic Therapy. Chest. 2004;126(3 Suppl):188S–203S. with molecularly imprinted polymer. Sensors Actuators B Chem.
82. Francis JL, Groce JB. Challenges in variation and respon- 2018;259:455–62.
siveness of unfractionated heparin. Pharmacotherapy. 94. Barzu T, Molho P, Tobelem G, Petitou M, Caen J. Binding and
2004;24(8P2):108S–19S. endocytosis of heparin by human endothelial cells in culture.
83. Kuhle S, Eulmesekian P, Kavanagh B, Massicotte P, Vegh P, Lau Biochim Biophys Acta. 1985;845(2):196–203.
A, et al. Lack of correlation between heparin dose and standard 95. Jaques L, Napke E, Levy S. The metachromatic activity of urine
clinical monitoring tests in treatment with unfractionated heparin following the injection of heparin. Circ Res. 1953;1(4):321–30.
in critically ill children. Haematologica. 2007;92(4):554–7. 96. Bjornsson TD, Wolfram KM, Kitchell BB. Heparin kinet-
84. Moynihan K, Johnson K, Straney L, Stocker C, Anderson B, ics determined by three assay methods. Clin Pharmacol Ther.
Venugopal P, et al. Coagulation monitoring correlation with 1982;31(1):104–13.
heparin dose in pediatric extracorporeal life support. Perfusion. 97. Bjornsson TD, Levy G. Pharmacokinetics of heparin. II.
2017;32(8):675–85. Studies of time dependence in rats. J Pharmacol Exp Ther.
85. Mann KG, Orfeo T, Butenas S, Undas A, Brummel-Ziedins K. 1979;210(2):243–6.
Blood coagulation dynamics in haemostasis. Hamostaseologie. 98. Boneu B, Caranobe C, Sie P. Pharmacokinetics of heparin
2009;29(1):7–16. and low molecular weight heparin. Baillieres Clin Haematol.
86. Newall F. Protamine titration. Methods Mol Biol. 1990;3(3):531–44.
2013;992:279–87. 99. McAvoy TJ. Pharmacokinetic modeling of heparin and its clini-
87. Crowther MA, Berry LR, Monagle PT, Chan AK. Mechanisms cal implications. J Pharmacokinet Biopharm. 1979;7(4):331–54.
responsible for the failure of protamine to inactivate low-molec- 100. Edward Conrad H. Heparin-binding proteins in hemostasis,
ular-weight heparin. Br J Haematol. 2002;116(1):178–86. Chapter 8. In: Edward Conrad H, editor. Heparin-binding pro-
88. Ramamurthy N, Baliga N, Wakefield TW, Andrews PC, Yang teins. San Diego: Academic Press; 1998. p. 239–300.
VC, Meyerhoff ME. Determination of low-molecular-weight 101. Young E, Prins M, Levine MN, Hirsh J. Heparin binding to
heparins and their binding to protamine and a protamine analog plasma proteins, an important mechanism for heparin resistance.
using polyion-sensitive membrane electrodes. Anal Biochem. Thromb Haemost. 1992;67(6):639–43.
1999;266(1):116–24. 102. Lijnen HR, Hoylaerts M, Collen D. Heparin binding proper-
89. Ignjatovic V, Summerhayes R, Gan A, Than J, Chan A, ties of human histidine-rich glycoprotein. Mechanism and
Cochrane A, et al. Monitoring unfractionated heparin (UFH) role in the neutralization of heparin in plasma. J Biol Chem.
therapy: which anti factor Xa assay is appropriate? Thromb Res. 1983;258(6):3803–8.
2007;120(3):347–51.
90. Bromfield SM, Barnard A, Posocco P, Fermeglia M, Pricl S,
Smith DK. Mallard blue: a high-affinity selective heparin sensor