Hereditary angioedema (HAE) is characterized by recurrent mucosal and cutaneous swellings,
resulting from excessive bradykinin generation, which is the end product of the kallikrein/kinin
system.[1]
[2] HAE predominantly occurs in patients with congenital C1-inhibitor deficiency. C1-inhibitor
controls the activation of factor XII (FXII) and plasma prekallikrein (PK). Notably,
bradykinin-mediated HAE is also described in patients with normal levels and functionality
of C1-inhibitor.[3] Inhibition of PK is currently being investigated as a prophylactic treatment in
HAE. Activated FXII converts PK into plasma kallikrein (PKa), which cleaves high-molecular-weight
kininogen (HK) thereby liberating bradykinin. PKa can activate FXII, plasminogen,
and urokinase-type plasminogen activator.[4] During HAE attacks, both the coagulation cascade and fibrinolytic system seem to
be activated as evidenced by elevated prothrombin fragment 1 + 2 and D-dimer levels.[5] Either way, HAE patients do not have a prothrombotic tendency.[6]
[Fig. 1] gives an overview of the functions of PKa in the kallikrein/kinin, intrinsic coagulation,
and fibrinolysis systems. Congenital PK deficiency is a very rare condition that is
presumed to be asymptomatic, but has been linked to increased risk of thrombotic events.[7]
[8]
[9]
[10] Congenital PK deficiency is usually detected when coagulation assays are performed,
as the activated partial thromboplastin time is prolonged in the absence of PK.[6]
[11] The critical roles PK play in the kallikrein/kinin system and in in vitro coagulation
are well known, but paradoxically PK does not contribute to in vivo hemostasis.[12] The link between reduced levels of PK and thrombotic risk is less well established.
The validity of the claim that PK deficiency increases thrombotic risks stated in
the previously mentioned case series is hampered by the lack of adequate control groups
and the risks of both selection and publication biases. It is important to note that
the majority of individuals with PK deficiency are presumed to go unrecognized given
its largely asymptomatic nature. Thus, the occurrence of cardiovascular or thrombotic
events in these subjects may have accounted for coagulation assays that revealed rare
observations, which are then more likely to be published. Animal studies contradict
the hypothesis of increased thrombotic risk in the absence of PK.[13]
[14]
[15] However, if an increased thrombotic risk of PK deficiency does exist, this can be
caused by either enhanced clot formation via the intrinsic coagulation cascade or
via decreased fibrinolytic activity. Increased thrombin generation could theoretically
be caused by enhanced intrinsic factor XI (FXI) activation due to increased binding
to HK, as the latter protein is less bound to PKa in this scenario.[16] Elevated FXI levels are associated with an increased risk of arterial and venous
thromboses.[17]
[18]
[19]
[20] However, KLKB1−/−
mice have similar FXI plasma levels compared with wild-type mice.[13]
Fig. 1 Overview of the functions of plasma kallikrein in the kallikrein/kinin, intrinsic
coagulation, and fibrinolysis systems. Activated factor XII (FXIIa) converts plasma
prekallikrein (PK) into plasma kallikrein (PKa). PKa cleaves high-molecular-weight
kininogen (HK), resulting in cleaved HK (cHK) and bradykinin (BK). BK binds to its
receptor on endothelial cells, leading to vascular leakage and thus angioedema. PKa
also activates factor XII (FXII) which activates the intrinsic coagulation cascade,
starting with factor XI (FXI), which becomes activated FXI (FXIa). FXIa converts factor
X (FX) into activated FX (FXa), which converts prothrombin into thrombin and the residual
product prothrombin fragment 1 + 2. Thrombin converts fibrinogen into fibrin, which
forms blood clots together with blood cells and platelets. Thrombin is inhibited by
the formation of complexes with antithrombin (TAT) and activates thrombin-activatable
fibrinolysis inhibitor (TAFIa), which protects fibrin from being degraded by the fibrinolytic
system. PKa also activates urokinase-type plasminogen activator (uPA), which, together
with tissue-type plasminogen activator (tPA), is inhibited by plasminogen activator
inhibitor-1 (PAI-1). Both plasminogen activators convert plasminogen into plasmin.
Finally, PKa activates plasminogen as well. Plasmin degrades fibrin thereby releasing
D-dimers. Plasmin is inhibited by α2-antiplasmin thereby generating plasmin–α2-antiplasmin
complexes (PAP).
In this letter, we present the results of coagulation and fibrinolytic activity assays
in samples obtained from a phase 2 trial in 22 HAE patients before and after 4 months
(hereinafter referred to as follow-up) of treatment with either 80 mg PK antisense
oligonucleotides (donidalorsen) or placebo.[1] At follow-up the median percent reduction from baseline in the donidalorsen group
in PK levels was 75% (range: 36–94%) and the silica-based activated partial thromboplastin
times remained within reference values. Coagulation and fibrinolytic activity markers
at follow-up were compared with baseline and placebo-treated patients and the results
are shown in [Table 1]. All generic and specific assays measured at follow-up in the donidalorsen group
were comparable with baseline. None of the concentrations or activities in the donidalorsen
group differed significantly from the placebo group after Holm–Bonferroni correction.
A sensitivity analysis of all outcome parameters in the randomized study population
did not reveal any statistically significant changes compared with baseline or placebo
either.
Table 1
Coagulation and fibrinolysis markers
|
Laboratory tests
|
Reference values
|
Donidalorsen baseline
|
Donidalorsen follow-up
|
Delta donidalorsen (follow-up minus baseline)
|
p-Value comparison, donidalorsen baseline and follow-up
|
Placebo baseline
|
Placebo follow-up
|
Delta placebo (follow-up minus baseline)
|
p-Value comparison, delta donidalorsen and delta placebo
|
|
CT lag time (min),
mean (SD)
|
1.5–3.2
|
3.0 (0.6)
|
3.0 (0.6)
|
−0.0 (0.3)
|
1.00
|
3.4 (0.7)
|
3.1 (0.6)
|
−0.2 (0.5)
|
1.00
|
|
CT peak thrombin (%),
mean (SD)
|
63–154
|
109 (34)
|
129 (32)
|
21 (27)
|
0.24
|
96 (22)
|
112 (27)
|
8 (15)
|
1.00
|
|
CT ETP (%),
mean (SD)
|
61–146
|
98 (16)
|
108 (23)
|
9 (20)
|
1.00
|
104 (30)
|
107 (13)
|
1 (21)
|
1.00
|
|
Factor XI activity (%),
mean (SD)
|
67–149
|
110 (31)
|
115 (27)
|
5 (20)
|
1.00
|
105 (28)
|
110 (27)
|
5 (21)
|
1.00
|
|
Prothrombin fragment 1 + 2 (pMol/L),
median (IQR)
|
53–271
|
251 (186, 406)
|
161 (119, 259)
|
−71 (−183, −23)
|
0.64
|
225 (177, 267)
|
189 (163, 267)
|
−8 (−64, 5)
|
1.00
|
|
TAT complexes (µg/L),
median (IQR)
|
<4.6
|
2.6 (2.2, 5.6)
|
1.9 (1.7, 2.3)
|
−0.2 (−4.1, 0.4)
|
1.00
|
2.3 (2.0, 2.9)
|
2.3 (2.0, 3.2)
|
0.0 (−0.1, 0.1)
|
1.00
|
|
HK activity (%),
median (IQR)
|
60–130
|
87 (78, 97)
|
93 (84, 105)
|
7 (0, 15)
|
0.42
|
101 (84, 120)
|
92 (82, 110)
|
6 (−9, 16)
|
1.00
|
|
nCLT (%),
median (IQR)
|
50–150
|
87 (79, 96)
|
93 (82, 108)
|
6 (1, 10)
|
1.00
|
87 (74, 106)
|
94 (70, 111)
|
−6 (−8, 0)
|
1.00
|
|
D-dimer (µg/mL),
median (IQR)
|
<0.50
|
0.65 (0.44, 2.80)
|
0.36 (0.27, 0.55)
|
−0.20 (−2.25, −0.03)
|
0.40
|
0.36 (0.36, 0.68)
|
0.52 (0.43, 0.54)
|
0.10 (0.01, 0.16)
|
0.64
|
|
Plasminogen activity (%),
mean (SD)
|
75–150
|
116 (25)
|
125 (21)
|
8 (16)
|
1.00
|
107 (23)
|
109 (25)
|
3 (19)
|
1.00
|
|
PAP complexes (ng/mL),
median (IQR)
|
0–514
|
533 (276, 1,156)
|
328 (187, 381)
|
−75 (−741, 7)
|
0.79
|
237 (201, 328)
|
212 (191, 257)
|
−24 (−73, 16)
|
1.00
|
|
α2-antiplasmin activity (%),
mean (SD)
|
80–120
|
122 (10)
|
127 (10)
|
7 (10)
|
0.41
|
116 (16)
|
121 (11)
|
5 (12)
|
1.00
|
Abbreviations: CT, calibrated automated thrombogram; HK, high-molecular-weight kininogen;
IQR, interquartile range; nCLT, normalized clot lysis time; PAP, plasmin–α2-antiplasmin;
SD, standard deviation; TAT, thrombin–antithrombin.
Note: Assessments at baseline and after 16 weeks of monthly treatment with 80 mg donidalorsen
or placebo. p-Values are adjusted for multiple comparisons with the Holm–Bonferroni method. n is 16 for all donidalorsen assessments except for baseline nCLT and follow-up TAT
where n is 14, and baseline α2-antiplasmin and D-dimer and all baseline CT parameters where
n is 15. n is 6 for all placebo assessments except for all baseline CT parameters and baseline
nCLT where n is 5.
We conclude that partial PK reduction of approximately 75% in HAE patients, albeit
very effective in reducing attack frequency, does not translate into an increased
coagulation activity or a decreased fibrinolytic activity. From these observations,
one could infer that the thrombotic risk is not increased with PK deficiency. Our
results are consistent with in vivo experiments in mice, as well as findings in HAE
patients treated with lanadelumab.[13]
[14]
[15]
[21] The contrast in our findings with the suggested increased thrombotic risk postulated
by Girolami et al and Barco et al[7]
[8]
[9]
[10] may be explained by methodological limitations or the more pronounced decrease of
PK levels in patients with a congenital deficiency. We did not observe a decrease
in HK activity after targeted PK reduction. This has not been previously investigated,
but conversely a congenital HK deficiency has been reported to be associated with
lower PK levels.[22]
[23] Binding to HK may protect PKa from inactivation or clearance.[24]
[25] Alternatively, congenital HK deficiency may be genetically linked to inherited PK
deficiency. Our study has several notable strengths. We made paired comparisons between
baseline and follow-up measurements in the same individuals. Additionally, we compared
our results to an adequate control group of patients with the same condition treated
with placebo. Another benefit is that we assessed both global measurements of the
thrombin-forming and fibrinolytic systems, as well as a more detailed examination
of several crucial enzymes and complexes within these systems. A limitation due to
the restricted use of remaining plasma samples from a previously completed trial was
the insufficient material available for all analyses in all patients. The samples
which lacked adequate data were missing at random and are thus not expected to have
influenced the validity of our results. A potential limitation of this study was the
approximately 75% reductions in PK levels. This has previously been shown to be sufficient
for decreasing angioedema attack rates[1] and we contributed with our study to the growing number of evidence that this amount
of PK reduction does not increase thrombotic risk in HAE patients. However, we acknowledge
that in congenital PK deficiency the PK levels are lower than in our study, meaning
that our results cannot be extrapolated one-to-one to the thrombotic risk in patients
with congenital PK deficiency.
In summary, our results do not demonstrate a procoagulant state in patients with approximately
75% reduced PK levels. This questions the earlier reported link between PK deficiency
and increased thrombotic risk. In addition, we showed that HK activity, FXI activity,
and plasminogen activity are not hampered by significantly reduced PK levels.
