Novel interactions between UFH and TFPI in children


Fiona Newall, Clinical Haematology Department, Royal Children’s Hospital, Flemington Rd, Parkville Victoria 3052, Australia.


The impact of age upon therapeutic response to unfractionated heparin (UFH) in children is proposed to reflect quantitative and potentially qualitative differences in coagulation proteins across childhood. This study explores the UFH-dependent tissue factor pathway inhibitor (TFPI) release in children compared to previously published data in adults. Children <16 years of age undergoing cardiac angiography formed the population for this prospective cohort study. TFPI release was measured prior to (baseline) and at 15, 30, 45 and 120 min post-UFH dose. This study demonstrated that, whilst the immediate release of TFPI post-UFH was similar in children compared to adults, TFPI release in children remained increased and consistent for a significantly longer period post-UFH administration compared to adults. Plasma TFPI levels in children did not demonstrate an UFH concentration –dependent reduction, as has been previously reported in adults. The prolonged TFPI-mediated anticoagulant levels observed in children administered UFH may contribute to the increased rate of major bleeding reported in children compared to adults. Furthermore, we postulate that this sustained UFH-dependent increase in TFPI levels in children may influence the binding of UFH to competitive plasma proteins, such as those involved in the immunological response to UFH associated with heparin-induced thrombocytopenia.

A reported ‘post-heparin effect’ was first described in 1963 when it was observed that the unfractionated heparin (UFH) antagonist, polybrene, was unable to completely reverse the in vivo effect of UFH(Nordoy, 1963). This ‘post-heparin effect’ has since been attributed to the impact of tissue factor pathway inhibitor (TFPI) released following UFH administration (Broze (Jr), et al (1988), Lindahl, 1994; Lindahl et al, 1991a,b, 1992; Sandset, 1996; Sandset & Abildgaard, 1991; Sandset et al, 1988, 2000, 1991). Three to 10 min after exposure to intravenous UFH, the concentration of free TFPI in adults increases 8- to 15-fold compared to baseline levels, before decreasing proportionally with the decrease in UFH concentration (Sandset et al, 2000). The in vivo anticoagulant effect of TFPI is twofold: first, TFPI binds to activated factor X (FXa via one of its three Kunitz-type domains, causing direct inhibition of FXa activity; second, after binding to FXa, a second Kunitz domain on the TFPI molecule binds to the previously formed complex of tissue factor (TF) and FVIIa. This quaternary (FXa-TFPI-TF-FVIIa) complex neutralizes the pro-coagulant catalytic effect of the TF:FVIIa complex(Sandset & Abildgaard, 1991; Ostergaard et al, 1993; Brodin et al, 2004). UFH is believed to increase the antithrombotic effect of TFPI in a dose-dependent fashion by increasing the affinity of TFPI for FXa by simultaneously binding to both proteins(Huang et al, 1993). Between one-third and one half of the FIIa inhibition achieved following the administration of UFH in vivo is attributable to TFPI(Lindahl et al, 1991b), with TFPI-potentiated inhibition of FXa increasing threefold.

Research investigating the known paediatric-specific mechanism of action of UFH in children compared to adults has predominantly focused upon antithrombin (AT), FIIa and FXa (Chan et al, 2002, 2003; Ignjatovic et al, 2006a,b, 2007; Monagle et al, 2008; Newall et al, 2009a,b, 2010). Whilst these studies have shed much light upon the management of UFH therapy in children, the contribution of TFPI to the differences observed in UFH response in children compared to adults has not been previously explored. This study aimed to determine, for the first time, the impact of UFH upon TFPI release in a population of infants and children receiving a single bolus dose of UFH.


Infants and children requiring cardiac angiography procedures between October 2006 and November 2008 at the Royal Children’s Hospital in Melbourne formed the participant pool in this study. Exclusion criteria were:

  • 1 Concomitant aspirin or non-steroidal anti-inflammatory therapy in the 10 d preceding the scheduled angiography procedure and/or concomitant anticoagulant therapy.
  • 2 A serum haematocrit result >0·53.

An intravenous catheter (IV) was placed for study purposes in the left or right long saphenous vein of participants following induction of anaesthesia. From this catheter a baseline, pre-UFH blood sample was collected. Further blood samples were collected at 15, 30 and 45 min post-UFH administration. A final blood sample was collected at 120 min post-UFH, or prior to the removal of the IV at the time of discharge from the Operating Theatre Recovery Room. All samples were collected into S-Monovette® tubes (Sarstedt Technology Park, S.A., Australia), containing 0·106 (3·2%) mol/l trisodium citrate anticoagulant in a ratio of nine volumes of blood to one volume of anticoagulant The samples were centrifuged at 1000 g for 10 min, at 10°C and plasma was separated into 250 μl aliquots before being stored at −85°C for batch testing.

TFPI levels were measured using the ASSERACHROM free TFPI enzyme-linked immunosorbent assay kit (Diagnostica Stago, Asnieres, France). The concentration of UFH in plasma was measured by protamine titration using a method adapted from Refn and Vestergaard (1954) to enable performance of the protamine titration assay using 100 μl of plasma (Newall et al, 2008).

Clinical data collected included the participant’s age, gender, weight, height and underlying cardiac anomaly. The study received institutional ethics approval (EHRC 26065A) and written informed consent was obtained from all study participants.

Statistical software package stata, Release 10.0 (STATA Corporation, College station, Texas, USA) was utilised for data processing and analysis. Numerical variables were presented as means and standard deviations.


Fifty-five participants (48% males) were recruited for this study. Their mean age was 5·7 ± 3·6 years (range 6 months to 15·2 years) with a mean weight of 20·3 ± 11·4 kg. An underlying diagnosis of patent ductus arteriosis was the most common indication for angiography (31%), followed by atrial septal defect (16%) and previous cardiac transplantation (16%). Seventy-three percent of patients had an interventional angiographic procedure, including balloon dilatation and device implantation. All remaining procedures were diagnostic. Eight patients met the institutional criteria for cyanotic heart disease, and had serum haematocrit levels >0·41. The mean dose of UFH administered to the study population was 96·2 ± 1·1 iu/kg.

The mean TFPI concentration for the 56 participants prior to UFH bolus was 11·3 ± 4 μg/l, with the results summarised in Table I. Following UFH administration there was a significant increase in TFPI release from the vascular endothelium. Specifically, for 86% of samples TFPI levels were greater than the limits of detection for this assay (i.e. >126·4 μg/l). Measurable TFPI levels were decreased by 33% at 102 ± 25 min post-UFH bolus compared to 17 ± 6 min post-UFH. The age-related differences across the four age-groups could not be determined.

Table I.   TFPI and protamine titration assay results according to time post-UFH bolus. Results are presented as the total number of tests or mean ± SD.
 Time post-unfractionated heparin Bolus (minutes)
Baseline17 ± 630 ± 347 ± 8102 ± 25
  1. *Results beyond the limits of detection of the commercial assay.

Tissue factor pathway inhibitor
 Number of recordable results4811419
 Mean ± SD (μg/l)11·3 ± 4124·8102·8114·8 ± 2·782·9 ± 10·1
 Number of unrecordable results (>126·4 μg/l)*044434215
Protamine titration
 Mean ± SD (units/ml)n/a2·2 ± 0·71·8 ± 0·81·4 ± 0·70·7 ± 0·3

The plasma concentrations of UFH, as measured by protamine titration, are presented in Table I. The concentration of UFH as measured by protamine titration decreased by 68% at 102 ± 25 min post-UFH bolus compared to 17 ± 6 min post-UFH.


This study investigated the effect of UFH upon TFPI release in a paediatric population receiving a single bolus dose of UFH for primary thromboprophylaxis, in the setting of cardiac angiography. In adults, UFH-mediated inhibition of FIIa and FXa via TFPI is short acting, with TFPI levels decreasing in response to the decreasing concentration of UFH(Sandset et al, 2000). This paediatric-specific study demonstrated that administration of UFH to children induces such a sustained level of TFPI that, for 86% of study participants, TFPI values exceeded the limits of detection of the assay for up to 2 h post-UFH bolus. The analysis of age-related differences in TFPI release following UFH was impeded by the high proportion of samples with TFPI levels above the limits of detection of the TFPI assay kit used in this study. This was despite an additional dilution step from 1:10 to 1:20 in an attempt to measure the actual concentration of TFPI in plasma. This mechanism of UFH action in children is reported for the first time and highlights the fact that patient age can significantly impact the physiological response to medications.

For the children in this study, TFPI levels did not decrease in response to the decreasing concentration of UFH following bolus dose administration. It has been previously reported that TFPI is responsible for between one-third and one half of the anticoagulant effect induced by UFH (Lindahl et al, 1991a,b), an effect that is not neutralized by polybrene or protamine(Abildgaard, 1992, 1993). Therefore, the sustained and prolonged TFPI release in children observed in our study compared to published adult data may contribute significantly to the increased rate of major UFH-related bleeding events observed in children compared to adults(Hull et al, 1990; Kuhle et al, 2007). The clinical impact of the sustained TFPI-mediated anticoagulant effect of UFH upon the risk of haemorrhage may be further exacerbated by the fact that this effect cannot be reversed by administration of protamine.

The possible mechanisms that may contribute to this age-dependent difference in UFH-mediated TFPI release are proposed here. First, infants and children may have higher physiological concentrations of TFPI within the vascular endothelium or intracellular stores compared to adults. Such a hypothesis is supported by the relative thrombo-protective characteristic of children compared to adults(Andrew et al, 2000), despite the decreased circulating total and free TFPI in healthy children compared to adults(Monagle et al, 2006). Second, there may be an age-related variability in TFPI clearance in children compared to adults. This is consistent with evidence arising from studies of UFH in neonates and children, suggesting UFH has an age-specific pharmacokinetic profile in infants and children compared to adults(McDonald et al, 1981; Andrew et al, 1988a; Schmidt et al, 1989). As an extension of this hypothesis, clearance of UFH-bound TFPI may therefore similarly vary with age. Third, the structure of the TFPI molecule could change with age, such that the affinity of the binding between UFH and TFPI could be increased in children compared to adults. Qualitative changes in coagulation proteins with age have been previously proposed as potential contributors to the age-related differences in UFH response observed in children(Newall et al, 2009b). Whilst there is indisputable evidence supporting that the concentration of serine proteases changes with age across infancy and childhood (Andrew et al, 1992; Andrew, 1995; Monagle et al, 2006), to date, our understanding regarding the impact of age upon serine protease structure remains limited. Studies using animal models and human plasma have demonstrated that fibrinogen exists in two isoforms, a fetal and an adult form (Witt et al, 1969; Witt & Muller, 1970; Francis & Armstrong, 1982; Hamulyak et al, 1983; Andrew et al, 1988b,c), whilst the existence of two isoforms of AT in the fetal plasma of sheep has also been reported (Niessen et al, 1996).

For the first time, we hypothesize that the prolonged UFH-induced TFPI release in children may contribute to the decreased incidence of heparin-induced thrombocytopenia (HIT) in children compared to adults (Newall et al, 2003, 2009a). Specifically, the activation of the haemostatic system associated with HIT that is known to contribute to increased thrombin generation and risk for thrombosis may be prevented by the increased circulating TFPI in children in the presence of high doses of UFH. Identification of the factors that protect children from developing HIT following UFH exposure may afford significant benefit to adults, where the morbidity and mortality of this UFH-related adverse event is so significant.

This study aimed to determine whether TFPI levels in children exposed to UFH could provide further insights into the mechanism of UFH action in children. Unfortunately, as the majority of TFPI measurements across all time-points post-UFH bolus were beyond the limits of detection for the TFPI method used, conclusions specific to an age-dependent mechanism of UFH response involving TFPI release across childhood were not possible.

Nonetheless, this study demonstrated for the first time that TFPI release following UFH administration in children is significantly prolonged compared to that previously reported in adults. Further investigation of the age-specific impact of UFH upon TFPI may provide significant insights into optimisation of UFH management in children. Furthermore, this research may contribute to an improved understanding of the variable adverse-event profile of UFH in children compared to adults with respect to major bleeding and HIT.


This project was supported by an NHMRC Research Grant #454385. Dr. Ignjatovic was supported by the Sanofi-Aventis Clinical Research Fellowship. Dr. Newall held an Australian Postgraduate Award.