Article, Pharmacology

Comparison of tranexamic acid plasma concentrations when administered via intraosseous and intravenous routes

a b s t r a c t

Introduction: There is a lack of information regarding intraosseous (IO) administration of Tranexamic acid (TXA). Our hypothesis was that a single bolus IO injection of TXA will have a similar pharmacokinetic profile to TXA administered at the same dose IV.

Methods: Sixteen male Landrace cross swine (mean body weight 27.6 +- 2.6 kg) were divided into an IV group (n = 8) and an IO group (n = 8). Each animal received 30 mg/kg TXA via an IV or IO catheter, respectively. Jug- ular blood samples were collected for pharmacokinetic analysis over a 3 h period. The maximum TXA plasma concentration (C_max) and corresponding time as well as distribution half-life, elimination half-life, area under the curve, plasma clearance and volume of distribution were calculated. One- and two-way analysis of variance for repeated measures (time, group) with Tukey’s and Bonferonni post hoc tests were used to compare TXA plasma concentrations within and between groups, respectively.

Results: Plasma concentrations of TXA were significantly higher (p b 0.0001) in the IV group during the TXA infusion. C_max occurred at 4 min after initiation of the bolus in the IV group (9.36 +- 3.20 ng/ul) and at 5 min after initiation of the bolus in the IO group (4.46 +- 0.49 ng/ul). Plasma concentrations were very similar from the completion of injection onwards. There were no significant differences between the two administration routes for any other pharmacokinetic variables measured.

Conclusion: The results of this study support pharmacokinetic bioequivalence of IO and IV administration of TXA.

Introduction

Following promising preliminary results from human studies such as the CRASH-2 and MATTERs studies, administration of intravenous

(IV) Tranexamic acid , an anti-fibrinolytic agent, is currently in- cluded in several civilian and military human trauma and Resuscitation protocols [1,2]. Military Damage-Control Resuscitation Clinical Practice Guidelines state that the early use of IV TXA should be strongly consid- ered for any patient requiring blood products in the treatment of combat-related hemorrhage and is most strongly advocated in patients judged likely to require massive transfusion [2,3]. Although these stud- ies suggest that early administration of TXA (<= 1 h following trauma) is associated with the greatest decrease in hemorrhage-related deaths, there is also evidence to suggest that its administration >= 3 h after injury may be harmful [1]. The narrow therapeutic timeline for TXA

? Funding: Public Works Canada, Defence research and development Canada (W7702- 145634/A).

* Corresponding author.

E-mail address: srboysen@ucalgary.ca (S.R. Boysen).

underscores the need to establish effective routes of administration to ensure trauma patients receive TXA in a timely fashion.

Several current guidelines state that if resuscitation is required and IV access is not obtainable, the intraosseous route should be used for certain medications [3,4]. Given the safety of administration of other medications via the IO route, and the fact that TXA has a pH similar to that of sodium chloride 0.9% solution and blood products [5], the use of IO TXA has been suggested should IV access not be available [4,6]. Despite evidence supporting the IO administration of drugs, there is a paucity of literature regarding the administration of IO TXA as a life- saving option in cases of failed or impossible IV catheterization [7].

The limited use of TXA via the IO route is likely explained by the fact that it has not been established if alternative routes of administration of TXA (i.e., IO, intramuscular, etc.) provide the same benefits that have been prospectively demonstrated with the IV route [5]. This has led to some authors cautioning against the use of IO TXA as equivalency between IO and IV routes has not been established and Pfizer recommends only intravenous administration, for which the detailed pharmacokinetics of TXA have been demonstrated [8]. It is apparent that studies investigating the administration of IO TXA are needed. In fact, on the initiative of the American military, a group of panel experts

http://dx.doi.org/10.1016/j.ajem.2016.10.054

recently identified potential routes of TXA administration (IV, oral, IO, transmucosal) as a Priority 2 research requirement [5].

The objective of this study was to determine if plasma levels of TXA are similar when given IO versus IV. We hypothesised that a single bolus IO injection of TXA will have a similar pharmacokinetic profile to TXA administered at the same dose IV.

Methods

This study was approved by the University of Calgary Animal Care Committee (Protocol AC15-0068) and the animals were managed in ac- cordance with the Canadian Council on Animal Care standards.

Animal Preparation

Sixteen male Landrace-Large White cross swine with a mean body weight of 27.6 +- 2.6 kg were included in the study. The pigs were sourced from a commercial swine production unit, housed individually or in groups of 2 and fed a pelleted swine ration twice daily with ad libitum access to nipple water feeders. Each pig was kept in an individ- ual pen and fasted with ad libitum access to water for twelve hours be- fore premedication.

Anesthesia

Swine were premedicated with intramuscular azaperone 6 mg (Stresnil 40 mg/ml, Vetoquinol Inc., Lavaltrie, QC, Canada J5 T 3S5), dexmedetomidine 0.6 mg (Dexdomitor 0.5 mg/ml, Zoetis Inc., Kalama- zoo, MI, USA 49007) and alfaxalone 60 mg (Alfaxan 10 mg/ml, Jurox Pty Ltd., Rutherford, NSW, Australia 2320). Swine were weighed prior to induction of anesthesia. A 22 gauge (G), two-port catheter (BD Saf- T-Intima IV Catheter, Becton Dickinson, Sandy, UT, USA 84070) was placed in an auricular vein of both ears and induction was completed with 1-2 mg/kg IV alfaxalone, titrated to effect. Animals receiving IV TXA had one ear marked ‘TXA’ in permanent ink to ensure that that particular auricular catheter was dedicated to the TXA bolus only.

Anesthesia was delivered using a circle system with isoflurane (F_I‘Iso

1.3-2%) and oxygen (F_I‘O₂ 100%). Two constant rate infusions (alfaxalone 25 ug/kg/min and dexmedetomidine 2 ug/kg/h) were ad- ministered via one of the auricular catheters using digitally programma- ble syringe drivers (Medfusion 3500, Smiths Medical International, St. Paul, MN, USA 55112). A 7.5 ml/kg/h continuous rate infusion (CRI) of 0.9% saline (1000 ml 0.9% Sodium Chloride, Baxter Corp, Mississauga ON, Canada L4Z 3Y4) was administered for the duration of the experi- ment into the TXA-designated auricular catheter to maintain patency.

Instrumentation

Instrumentation included a 5-lead electrocardiogram (ECG), buccal Pulse oximeter, sidestream capnograph, spirometer, tissue oxygen satu- ration monitor, bispectral index (BIS) and a rectal thermometer. The ECG, pulse oximeter, capnograph and spirometry data were displayed and recorded with a multiparameter monitor (Datex-Ohmeda s/5 Col- lect, GE Healthcare, Helsinki, Finland FIN-00031). The BIS electrodes (BIS Pediatric(TM) Sensor, Covidien LLC, 15 Hampshire Street, Mansfield, MA, U.S.A. 02 048) were placed horizontally across a clipped area of skin overlying the frontal bone and right lateral aspect of the skull and connected to a BIS monitor (BIS Vista Monitoring System, Aspect Medical Systems, Norwood, MA, USA 02062). Animals were placed into dorsal recumbency for the duration of the study.

One Common carotid artery was catheterized to perform intermit- tent arterial blood gas analysis and continuous arterial Blood pressure monitoring (systolic, diastolic and mean). An external jugular vein was also catheterized to allow TXA blood sampling. A tissue saturation sensor (Inspectra StO₂ Sensor, model 1615, Hutchinson Technology, 40 West Highland Park Drive NE, Hutchinson, MN, USA 55350) was

placed over the right adductor muscle. Rectal body temperature was measured with a pliable rectal probe (Datex-Ohmeda s/5 Collect, GE Healthcare, Helsinki, Finland FIN-00031) and maintained between 37 and 38 ?C using a Circulating water pad as needed.

The IO TXA animals received an IO catheter (Arrow EZ-IO(R), pediatric, pink 15 G x 15 mm, Teleflex Incorporated, Wayne, PA) that was inserted by a single experienced operator into the right proximal tibia just medial to the tibial tuberosity. A power drill (Arrow EZ-IO(R), Power Driver – G3, Teleflex Incorporated, Wayne, PA) was used and a 90? angle to the bone surface was maintained during insertion. The stylet was removed, blood was observed back-flowing through the hub, and an extension set was attached (EZ(R) Connect, Teleflex Incorporated, Wayne, PA). A small amount of blood was aspirated and the catheter was flushed with 20 ml heparinized saline to ensure patency.

Experimental Design

Sixteen animals were used in the study and all were included in the data analysis. Animals were divided into two groups. Each pig in the IV group (n = 8) received 30 mg/kg of IV TXA (Tranexamic Acid 100 mg/ml, Sandoz, QC, Canada, J4B 7 K8), followed by a 1 ml/kg saline bolus. Each pig in the IO group (n = 8) received 30 mg/kg TXA via the IO catheter followed by a 1 ml/kg saline bolus. The TXA dose was made up to a total of 25 ml using saline as the diluent in both groups and the so- lution was administered over 5 min using a digitally programmable sy- ringe driver. The saline bolus was manually administered over 1 min.

None of the observers were blinded to treatment. The experimental protocol was divided into 2 phases. A detailed data collection timeline is shown in Table 1. The first 30 min established anesthetic and physiolog- ic equilibration prior to the second phase. During the second phase physiologic recording and blood sampling were conducted. Data includ- ing heart rate, ECG, direct arterial blood pressure, end tidal CO_2, BIS, tis- sue oxygen saturation, respiratory rate, tidal volume, minute volume, and rectal temperature were collected continuously. Arterial blood

Table 1

Timeline for pharmacokinetic, arterial blood gas, and physiologic data collection prior to and following intraosseous (IO, n = 8) and intravenous (IV, n = 8) tranexamic acid administration of 30 mg/kg over 5 min and a total jugular venous sampling period of 3 h. Intravenous injection was via an auricular vein and IO infusions were through the right proximal tibia.

Time point	Description	Data collected
SSA	Steady state anesthesia	pK, ABG, PP
Baseline (BL)	Immediately before TXA bolus	pK, ABG, PP
Ti	TXA bolus 1 min	pK, PP
Tii	TXA bolus 2 min	pK, PP
Tiii	TXA bolus 3 min	pK, PP
Tiv	TXA bolus 4 min	pK, PP
Tv	TXA bolus 5 min	pK, ABG, PP
T1	Post TXA bolus + 1 min, end saline bolus	pK, PP
T2	Post TXA bolus + 2 min	pK, PP
T4	Post TXA bolus + 4 min	pK, PP
T6	Post TXA bolus + 6 min	pK, PP
T8	Post TXA bolus + 8 min	pK, PP
T10	Post TXA bolus + 10 min	pK, ABG, PP
T15	Post TXA bolus + 15 min	pK, ABG, MVBG, PP
T20	Post TXA bolus + 20 min	pK, ABG, PP
T25	Post TXA bolus + 25 min	pK, ABG, PP
T30	Post TXA bolus + 30 min	pK, PP
T45	Post TXA bolus + 45 min	pK, ABG, MVBG, PP
T60	Post TXA bolus + 60 min	pK, ABG, PP
T80	Post TXA bolus + 80 min	pK, ABG, PP
T100	Post TXA bolus + 100 min	pK, ABG, PP
T120	Post TXA bolus + 120 min	pK, ABG, PP
T150	Post TXA bolus + 150 min	pK, ABG, PP
T180	Post TXA bolus + 180 min	pK, ABG, PP

PP: physiologic parameters including cardiac and respiratory values, tissue oxygen satura- tion, rectal body temperature, and bispectral index data were collected continuously; pK: jugular venous sample for pharmacokinetic analysis; ABG: arterial blood gas; MVBG: mixed venous blood gas.

samples were intermittently collected throughout the study. Jugular ve- nous blood samples were collected for pharmacokinetic (pK) analysis in the following sequence: during steady state anesthesia (SSA), at base- line (BL) immediately before TXA injection, every one minute during TXA injection (Ti-v), and 1, 2, 4, 6, 8, 10, 15, 20, 25, 30 45, 60, 80, 100,

120, 150 and 180 min (T1-180) after TXA injection (Table 1).

Each sample was either processed immediately or stored in an ice bath for a maximum of 10 min. A single operator processed all of the samples. Vacutainer tubes were centrifuged at 3600 RPM for 10 min at 4 ?C. A pipette was used to draw off the plasma and duplicate cryovials were filled with 0.5 ml plasma. The cryovials were submerged in liquid nitrogen, removed, and immediately placed into a – 80 ?C freezer.

Euthanasia and Pathological Examination

Upon completion of the experiment, isoflurane was increased to 4% for 10 min and animals were euthanized with 12 ml of a saturated KCL solution administered via IV injection. Following euthanasia, both pelvic limbs were disarticulated at the stifle joint in the IO group and the distal portion of the limbs were immediately placed on ice and submitted to a USA board-certified anatomic pathologist (CK) for gross and microscop- ic examination of the tibial medullary cavities.

Each tibia was clamped in a vice and a double-bladed saw was used to cut a 5 mm thick longitudinal slab that included the Proximal tibial epiphysis, the growth plate, and the proximal one third of the metaphysis. For the right tibia the slab included the catheter introduc- tion site; for the left tibia the slab mirrored that taken from the right side. Slabs were decalcified in 20% formic acid for 2-3 days until sufficiently softened to be trimmed with a scalpel blade. After decalcifi- cation, an approximately 20 x 30 mm section was cut from each slab, embedded in in paraffin, and prepared routinely for histologic examination using hematoxylin and eosin stain. Each histologic section included proximal tibial articular cartilage, epiphysis, growth plate and metaphysis.

Sections from the right tibial slab were cut in such a way that the catheter path within the metaphyseal cortex was avoided; the purpose of this was to assess potential bone marrow lesions caused by drug toxicity rather than by catheter insertion trauma. Bone marrow in each section was assessed microscopically in a blinded manner (left versus right side) according to previously published methods [9,10].

Determination of Plasma Tranexamic Acid Concentrations

The materials used included acetonitrile and methanol LC/MS grade (Sigma-Aldrich, Mississauga, ON, Canada L6H 6J8), formic acid ACS grade (EMD-Millipore Ltd., Toronto, ON, Canada L6H 6J8), TXA for IV injections preservative free (Tranexamic Acid 100 mg/ml, Sandoz, QC, Canada, J4B 7K8) and ultrapure water (18.2 M?). Plasma samples were thawed at 37 ?C for 10 min and vortexed. Proteins were precipitat- ed by mixing 100 ul of the plasma sample with 300 ul of ice cold methanol. Mixtures were incubated at -20 ?C for 20 min prior to cen- trifugation at 17 000 xg for 10 min. A 40 ul aliquot of the supernatant was added to 960 ul of methanol in 1 ml auto-sampler vials. These sam- ples were stored at -20 ?C prior to analysis. The analytical LC/MS for the quantification of TXA consisted of Agilent 1200 LC (binary pump, standard auto injector, and temperature-controlled column compart- ment) coupled with Agilent MS TOF 6230.

The separation was performed on an Agilent 1200 HPLC with a Zorbax eclipse XDB-C18 column. The column temperature was set at 40 ?C. The mobile phase consisted of Buffer A (10 mM formic acid in water) and Buffer B (10 mM formic acid in 90% acetonitrile). Injection volume was 1 ul. A linear gradient elution was performed at a flow rate of 0.4 ml/min, with gradient changes at 0 min of 3% Buffer B, at 8 min to 85% Buffer B, at 10 min to 100% Buffer B and at 12 min to 3% Buffer B.

The mass spectral analysis was performed on an Agilent 6230 TOF MS controlled by MassHunter software. The instrument was operated in positive mode using electron ionization (EI) source and following operational settings: gas temperature of 300 ?C, gas flow of 4 l/min, nebulizer at 50 psi, sheath gas temperature at 250 ?C, sheath gas flow at 10 l/min, nozzle voltage at 500 V and fragmentor voltage of 90 V. The limit of detection (LOD) in plasma was 0.01 ug/ml.

Pharmacokinetic Analysis

The pharmacokinetic analyses were performed with PK Solutions using a non-compartmental pK method. Plasma concentration versus sampling time data were plotted for each animal and Pharmacokinetic parameters were calculated. The maximum TXA plasma concentration for each animal (C_max) and its corresponding time (T_max) were reported directly from the assay data following administration by each route. Individual elimination rates (k_el) were determined by linear regression of the respective natural logarithm of TXA plasma concentrations versUS time during the terminal phase. The following variables were determined: distribution half-life (t_1/2 D), elimination half-life (t_1/2 E), area under the curve (AUC), plasma clearance (Cl_p; where Cl_p = Dose / AUC_(0-?)), and volume of distribution (V_d; where Vd = Dose / AUC_(0-?) x k_el).

Statistical Analyses

All data were reported as mean plus or minus one standard devia- tion (SD) unless otherwise stated. Plasma concentration graphs were plotted as mean plus or minus one standard error mean (SEM). Shapiro-Wilk analysis was used to test data for normality. Parametric analyses were used where data approximated normal distribution. Var- iables compared over time and between groups were analyzed with one and two-way analysis of variance for repeated measures (time, group) with Tukey’s and Bonferonni post hoc tests, respectively. Differences in body weight, drug dosage and pharmacokinetic parameters were an- alyzed with a Mann-Whitney U test. Derived pharmacokinetic compar- isons were plotted as median plus or minus interquartile range (IQR). Differences were considered significant at p <= 0.05. Statistical analyses were performed with GraphPad Prism(R) 5.04 (GraphPad Prism, La Jolla, CA, USA 92037).

Results

Data from all sixteen animals (n = 8 per group) were used in the study with the exception of the arterial blood gas analyses, where some data points were missing. There were a total of 5 ABG data points missing for time point Tv (2 in the IV group and 3 in the IO group) and all data for this time point were excluded from ABG statistical analysis. There were also missing ABG data points at T5 in one swine from the IV group, and at T120, T150, and T180 from another swine in the IV group. This resulted in only 6 swine being included in ABG statistical analysis in the IV group. There were no significant differ- ences in body weight or premedication doses of alfaxalone, azaperone or dexmedetomidine between the IO and IV groups (2.06 +/- 0.24,

0.21 +/- 0.02, 0.02 +/- 0.0, and 2.19 +/- 0.26, 0.22 +/- 0.02,

0.02 +/- 0.0 mg/kg, respectively). Mean body weight (+-SD) was

28.8 +- 2.4 and 26.5 +- 2.4 kg in the IO and IV groups, respectively. There was no significant difference between total dose of TXA administered (p = 0.08). Mean TXA doses (mg) were 863 +- 72 (IO) and 795 +- 72 (IV).

Catheter Placements

There were no complications associated with auricular venous or tibial IO catheterization. All IO catheters were inserted on the first attempt by the same investigator.

Physiologic Parameters

There were no significant differences between IO and IV groups for any cardiovascular parameters, rectal temperature, bispectral index or Tissue oxygenation. However, within each group there was a statistically significant variation in heart rate, arterial blood pressure, temperature, BIS and STO₂ over time. These changes were not clinically significant or acutely related to TXA administration. A sub-analysis was performed for mean heart rate as well as systolic, diastolic and mean arterial blood pressures over the five minute TXA injection period for both IO and IV groups. There were no significant differences over time or between groups for any of these parameters.

There were no significant differences between IO and IV groups for respiratory rate or arterial blood gas parameters. Within each group there were statistically significant variations in respiratory rate (n = 8 per group), and the following blood gas parameters over time (n = 6 IV, n = 8 IO); pH, partial pressure of arterial carbon dioxide, partial pres- sure of arterial oxygen, hemoglobin, sodium, arterial bicarbonate, and extracellular base excess (Table 2a, Table 2b). The changes were relatively small and considered clinically insignificant.

Pharmacokinetic Parameters

Pharmacokinetic data were collected and analyzed for all 16 animals. Changes in plasma concentration over time are shown in Fig. 1. Plasma concentrations of TXA were significantly higher (p b 0.0001) in the IV group during the TXA infusion (Figs. 1 & 2). Mean (+- SD) plasma TXA concentrations (ng/ul) over this interval were 7.25 +- 1.93 and 2.55 +- 1.61, respectively. Peak plasma concentrations (C_max) occurred 4 min after initiation of the bolus in the IV group (mean +- SD: 9.36 +- 3.20 ng/ul; range: 3.71-13.01 ng/ul) and at 5 min post-bolus in the IO group (mean +- SD: 4.46 +- 0.49 ng/ul; range: 3.63-5.13 ng/ul). Peak concentration for TXA was therefore 52% lower with IO administration compared to IV when a mean ratio was calculated (Table 3). Plasma concentrations were very similar from T1 (completion of the saline bolus post TXA administration) onwards (Fig. 1).

There were no significant differences between the two administra- tion routes for any other direct or calculated pharmacokinetic variable. Volume of distribution was moderately more variable in the IO group (Fig. 3). One IO pig also had significantly higher plasma clearance com- pared to all other animals (Fig. 4). Comparison of intraosseous/intrave- nous ratios of mean AUC_(0-?) values indicated that IO administration only resulted in a 5% lower exposure to TXA compared to the IV route (Fig. 5, Table 3).

Tibial Bone Marrow Pathology

Right and left tibias from all pigs in the IO group were examined. Bone marrow was histologically normal in all sections. There was no dif- ference in marrow between the right (catheterized) and left tibia for any animal.

Discussion

In emergency settings where patients are in shock with collapsed peripheral circulation, alternative routes such as IO catheters may be re- quired to improve patient outcome [11,12]. It is therefore essential to know if medications administered intraosseously can be given at the same dose and rate to achieve the same clinical benefit as the IV route. Although intravenous TXA has been used in various porcine experimen- tal models [13-15], the only documented use of IO TXA administration in pigs involved cadaver limbs [16]. The current study is the first pub- lished report comparing pharmacokinetic characteristics of IO and IV TXA in any species. The results of this study support bioequivalence of TXA when administered IO or IV in swine, and support previously de- scribed pharmacokinetic characteristics of IV administered TXA [17-19]. Given that studies suggest there is a limited therapeutic window in which the administration of TXA has a beneficial effect [1], the peak con- centration (C_max) and time to peak concentration (T_max) of TXA are clin- ically important to assess. In the current study, the IV route reached C_max 4 min after initiating the Bolus injection, which was 1 min earlier than the IO route. The disparity in TXA plasma concentrations and C_max dur- ing the five minute bolus period in this study was most likely due to the proximity of the auricular venous circulation versus the tibial circulation relative to the jugular catheter sampling site. The short lag time be- tween infusing the bolus and collecting serial venous samples likely did not allow for adequate vascular and corporeal mixing at the IV injec- tion site before reaching the sampling site [20]. Plasma concentrations during this interval in the IV group therefore most likely represented drug streaming and subsequent withdrawal of unmixed TXA. We hy- pothesis that mixed venous or arterial sampling sites may eliminate drug streaming if the auricular vein is used for TXA administration during TXA pharmacokinetic studies. Alternatively, if the jugular site is chosen for IV TXA pharmacokinetic study sampling, other IV drug ad- ministration sites that drain into the inferior vena cava (i.e. femoral vein), may allow sufficient mixing of TXA and blood to eliminate the risk of drug streaming. Further studies are required to confirm these hypotheses. The tibial IO catheter was likely a sufficient distance from the jugular circulation so the sampling intervals were inconsequential and adequate mixing occurred prior to blood collection. Therefore, it

Table 2a Mean (+-SD, underneath) physiologic and arterial blood gas parameters following porcine auricular intravenous (IV n = 6) bolus administration of 30 mg/kg tranexamic acid over 5 min at various time points during a jugular venous sampling period of 3 h.

	SSA	BL	Tv	T10	T15	T20	T30	T60	T80	T120	T150	T180
PH	7.37	7.38	7.37	7.37	7.38	7.38	7.39	7.39	7.40	7.41	7.40	7.40
	0.03	0.03	0.03	0.03	0.03	0.04	0.03	0.03	0.03	0.05	0.04	0.04
pCO₂ (mm Hg)	52.5	53.4	53.6	53.0	50.5	49.7	49.8	50.6	49.8	47.8	50.1	48.9
	6.6	7.2	7.3	9.0	5.6	4.8	5.2	6.2	4.8	7.0	5.8	6.0
pO₂ (mm Hg)	356	349	346	337	339	335	337	320	309	301	302	297
	31	31	32	37	39	66	46	45	54	74	76	74
Hgb g/L	110	111	111	109	109	110	110	109	110	109	108	108
	9	10	10	8	8	8	6	6	6	6.	6	7
Sodium (mmol/L)	140.5	139.9	140.6	140.8	140.6	140.3	139.7	138.9	139.0	139.3	139.4	139.1
	2.6	1.9	2.8	2.1	2.3	2.1	0.9	1.1	1.5	2.0	1.5	1.5
HCO₃ (umol/L)	30.3	31.2	31.0	30.7	29.6	29.6	29.8	30.3	30.5	29.8	30.6	30.1
	2.5	2.4	2.5	3.1	2.2	1.2	1.5	2.9	1.6	1.7	1.7	1.8
BE (umol/L)	4.7	5.4	5.2	4.4	4.6	4.6	4.7	5.3	5.7	4.8	5.9	5.3
	2.6	2.0	2.3	1.9	2.4	1.4	1.6	3.1	1.8	1.3	1.8	1.9
RR (bpm)	26.3	26.8	NA	28.00	NA	29.1	29.5	30.8	31.1	31.6	30.6	30.8
	5.8	5.751	NA	6.0	NA	6.1	6.0	6.3	6.0	5.9	5.3	6.5

pCO₂: partial pressure of carbon dioxide; pO₂: partial pressure of oxygen; Hgb: hemoglobin; HCO₃: bicarbonate; BE: base excess; RR: respiratory rate; bpm: breaths per minute; N/A: not assessed at that time point.

Table 2b

Mean (+-SD, underneath) physiologic and arterial blood gas parameters following porcine Tibial intraosseous (IO n = 8) bolus administration of 30 mg/kg tranexamic acid over 5 min at various time points during a jugular venous sampling period of 3 h.

	SSA	BL	Tv	T10	T15	T20	T30	T60	T80	T120	T150	T180
PH	7.38	7.38	7.38	7.38	7.38	7.39	7.39	7.40	7.4	7.4	7.40	7.40
	0.02	0.02	0.02	0.02	0.02	0.02	0.01	0.02	0.03	0.03	0.02	0.03
pCO₂ (mm Hg)	51.9	52.6	53.7	52.7	53.1	51.7	53.2	52.2	50.3	51.0	52.6	52.4
	4.7	5.5	5.0	4.1	4.6	4.9	4.2	3.3	4.8	5.8	4.6	5.6
pO₂ (mm Hg)	331	329	325	312	295	296.	293	288	281	283	268	265
	55	58	54.	69	75	79	69	665	63	67	79	74
Hgb g/L	110	110	109	107	109	106	108	106	106	106	106	105
	7	7	7	6	7	5	8	5	88	7	6	9
Sodium (mmol/L)	139.3	139.3	139.5	140.0	139.3	138.1	138.7	137.8	138.2	138.0	138.1	138.3
	1.1	1.0	1.5	1.1	1.8	2.0	1.8	1.6	1.7	1.5	1.4	1.1
HCO₃ (umol/L)	30.9	31.2	31.9	31.3	31.7	31.2	32.40	32.1	31.4	31.8	32.2	32.0
	1.9	2.2	1.7	1.4	1.8	1.9	2.6	1.7	2.5	2.2	1.9	1.9
BE (umol/L)	5.7	5.9	6.6	6.0	6.5	6.0	7.4	7.2	6.6	6.9	7.1	6.9
	2.0	2.1	1.7	1.3	1.9	1.9	2.7	1.8	2.6	2.1	1.8	2.1
RR (bpm)	26.3	26.8	NA	28.00	NA	29.1	29.5	30.8	31.1	31.6	30.6	30.8
	5.8	5.751	NA	6.0	NA	6.1	6.0	6.3	6.0	5.9	5.3	6.5

appears the proximal tibia is a good IO site to administer TXA for phar- macokinetic studies, and will likely allow sufficient mixing with a jugu- lar, mixed venous or arterial blood sampling sites.

Despite the statistically significant difference in C_max between the IV and IO routes, this is unlikely to result in a clinically significant differ- ence given the plasma concentrations were very similar in both groups within a minute of completing the injection, and a difference of 1 min to reach peak plasma concentration should not impact clinical outcome.

The pharmacokinetic results of our study are not surprising given dye and radioactive tracer studies indicate that drugs given via an IO catheter enter the central circulation within seconds, comparable to ad- ministration via peripheral IV catheters [21]. Several previous studies have demonstrated pharmacokinetic equivalence for a number of drugs given IO versus IV. For example, morphine (human), rocuronium (swine), atropine (swine) and epinephrine (swine, dogs) have demon- strated equivalent therapeutic plasma concentrations and pharmacoki- netic endpoints when administered intraosseously compared to intravenously [11,22-25]. Evidence also suggests there are no clinically relevant differences in drug kinetics between IO and IV-administered medications in hypo- and normovolemic states. However, most of

Fig. 1. Plasma concentration (mean +- SEM) versus time profiles (0 -3h) after intravenous (n = 8) and intraosseous (n = 8) bolus administration of tranexamic acid (TXA; 30 mg/kg over 5 min; Ti to Tv) in swine. Intravenous (IV) TXA administration was via an auricular vein. Intraosseous (IO) infusion was performed through the right proximal tibia. TXA was diluted to 25 ml with saline in both groups. BL: Baseline.

these studies were conducted using tracer dyes, and under very con- trolled hemorrhagic models, which may not equate to TXA or more se- vere states of Cardiovascular collapse [26,27].

Non-IV routes of administration for fluids and medications have been advocated in mass casualty settings [28] and IO administration is currently recommended as an alternative to IV drug administration in human cases of arrest [29,30]. The high first attempt success rate, rapid ability to place IO catheters and simplicity of newer IO devices make the IO route an attractive option in the prehospital and hospital setting for both fluid resuscitation and drug therapy [28,31-33]. Further- more, numerous studies have shown that the pharmacological effects of drugs, such as epinephrine on heart rate and blood pressure, are equiv- alent when given via either route [24,34].

Despite evidence to suggest medications can be given via the intraosseous route, the only documented case of IO TXA administration in humans was reported in a military setting [7]. A 2015 retrospective analysis of 1000 uses of IO access described IO delivery of TXA in 82 pa- tients [7]. All but 4 of these patients received IO TXA during helicopter evacuations performed by Defense Military Services in Iraq and Afghanistan. These casualties received IO catheters either as a primary choice in the case of blast-induced traumatic amputation or following

Fig. 2. Plasma concentration (mean +- SEM) versus time profiles during the interval of tranexamic acid infusion (30 mg/kg over 5 min; Ti to Tv) using intravenous (n = 8) and intraosseous (n = 8) routes of administration in swine. Plasma concentration changed significantly over time (?) in both groups (p b 0.0001). TXA plasma concentrations were significantly higher (*) in the intravenous (IV) group (p b 0.0001) compared to the intraosseous (IO) group at Ti (p b 0.01), Tii to Tiv (p b 0.0001) and Tv (p b 0.05). TXA was diluted to 25 ml with saline in both groups. BL: Baseline.

Table 3 Mean (+-SD) pharmacokinetic values following porcine tibial intraosseous (IO) and auric- ular intravenous (IV) bolus administration of 30 mg/kg tranexamic acid over 5 min and a total jugular venous sampling period of 3 h (IO n = 8, IV n = 8).

Parameter IO route (+-SD) IV route (+-SD) IO/IV AUC_(0-?) (ug-min/ml) 228.9 +- 51.1 240.6 +- 50.6 0.95

C_max (ng/ul)^?,? ?4.46 +- 0.49 ?9.36 +- 3.20 0.48

T_max (min)^? 5.00 4.00 1.25

Cl_p (ml/min/kg) 138.6 +- 39.7 129.2 +- 24.9 1.07

V_d (L/kg) 12.17 +- 2.2 12.20 +- 1.6 1.00

t_1/2 Dist (min) 4.0 +- 0.7 3.6 +- 1.0 1.10

t_1/2 Elim (min) 62.7 +- 11.1 66.5 +- 8.3 0.94

AUC: area under the curve, Cl_p: plasma clearance, Vd: volume of distribution, t_1/2 Dist: half-life of distribution, t_1/2 Elim: half-life of elimination.

* Denotes significant difference between groups (p = 0.007).

^? Disparity in these results likely due to a sampling error in the intravenous group – inadequate mixing time from auricular to jugular circulation resulted in partial withdraw- al of unmixed TXA.

Fig. 4. Plasma clearance (median +- IQR) of tranexamic acid (TXA) following a 30 mg/kg bolus over 5 min and a total jugular venous sampling period of 3 h. One pig in the IO group demonstrated a markedly higher clearance (227.2 ml/min/kg) compared to all other animals.

failure of peripheral catheterization. No complications were recorded and the survival rate was 69.2% [7].

The use of IO TXA has been questioned because there is no evidence of therapeutic equivalence between IO- and IV-administered TXA [8]. Furthermore, Pfizer recommends only IV administration, for which the pharmacokinetics of TXA have been demonstrated [8]. Our study contributes to the literature regarding the pharmacokinetics of intraosseously administered TXA and supports this route of administra- tion in swine models. This does not, however, equate to therapeutic equivalence and further studies are needed to determine if IO and IV TXA administration in hemorrhagic shock have the same therapeutic equivalence.

The pharmacokinetic characteristics of drugs delivered intraosseously can be altered by drug distribution to the bone marrow and blood flow to the bone [23]. A trend towards higher variability in the volume of distribution was observed when using IO administration in the current study although the difference between groups was not significant. Von Hoff et al. reported a significantly higher V_d in adults re- ceiving morphine intraosseously compared to intravenously through an IO cannula in the Iliac crest [11]. This finding was attributed to a small deposition effect near the IO port or in the bone marrow. The “depot” effect causes a drug to persist in the IO space, resulting in a lower peak concentration and a longer time to reach peak concentration [35]. It has been suggested that any depot effect will be greater when oil emulsions are used for drug delivery, as these emulsions are believed to remain in the marrow for longer periods creating a reservoir for drugs that are slowly liberated and dispensed by the oil [36]. A possible depot effect is unlikely significant in the current study as TXA is an aqueous solution, and flushing the IO catheter with fluids following medication administration is believed to reduce any potential “depot” effect [36].

Fig. 3. Volume of distribution (median +- IQR) of tranexamic acid (TXA) following a 30 mg/kg bolus over 5 min and a total jugular venous sampling period of 3 h. The IO group showed higher variability in volume of distribution compared to IV animals.

However, more hydrophilic drugs that distribute to the bone mar- row may create an absorption phase that requires dose titrations to maintain effective plasma concentrations [23]. An absorption phase due to bone marrow distribution following IO administration of TXA in the current study seems unlikely as the volume of distribution was not statistically different between groups and the IO plasma concentra- tion did not exceed the IV plasma concentration following completion of drug administration. As the current study only compared a single bolus injection of TXA, it is unclear if multiple boluses or a CRI of TXA would have produced similar IO and IV pharmacokinetic results. This should be investigated further.

A clinically significant side effect of IV TXA is hypotension as a result of rapid rates of infusion [37,38]. Injection of 30 mg/kg tranexamic acid (diluted to a total volume of 25 ml with 0.9% saline) given over 5 min, either intravenously or intraosseously, had no observable effect on mea- sured physiologic parameters in the current study. Changes in some physiologic variables over time most likely reflected mild cardiorespira- tory depression associated with prolonged general anesthesia and dor- sal recumbency. Decreased blood pressure has been reported in humans after TXA administration [39,40] but per minute data analysis during the TXA bolus interval showed no cardiovascular compromise. The dose and rate of TXA administration used in the current study did not result in hypotension or have any other detectable negative short term cardiovascular consequences. It is possible that a higher dose or rate of administration, repeated boluses, or a CRI would have resulted in hypotension during infusion of TXA.

The major limitation of this study was the use of a normovolemic, anesthetized animal model under Acute conditions. Although swine are considered one of the closest experimental equivalents to humans, these results may not translate to a pediatric or adult patient, particular- ly if they are in hemorrhagic shock. TXA is eliminated by glomerular

Fig. 5. Area under the curve (AUC) for intravenous (IV) and intraosseous (IO) tranexamic acid (TXA) administration. IO administration resulted in a 5% lower exposure to TXA compared to the IV route.

filtration, and excretion may be altered or delayed in trauma patients with compromised renal function [18].

Like many other pharmacologic studies, these findings are also sex- biased and only reflect male physiology. Human pharmacokinetic TXA data have only been reported in young adult men [17-19].

anatomical differences must also be considered. Tibial IO pharmaco- kinetics in immature swine may differ from a humeral or sternal site in adult humans. Humeral and sternal circulations are closer to central ve- nous circulation, and the sternal marrow compartment has a smaller volume compared to a long bone [36]. The IO TXA in this study was also injected into red marrow. Adult long bones contain almost exclu- sively yellow marrow after 20 years of age [36]. Yellow marrow has a different vascular structure but appears equally capable of delivering IO administered agents into the circulation.

Finally, the proximity of the auricular venous circulation to the jug- ular sampling site created a potential sampling artifact and may have skewed the C_max results. IO tranexamic acid C_max and T_max are potential- ly more equivalent to the IV route than this study demonstrated.

Conclusion

Although TXA is currently being administered intraosseously, there was no experimental evidence prior to this study to show that IO ad- ministration is pharmacologically and physiologically equivalent to IV administration. Results from this swine model support the pharmacoki- netic bioequivalence of IO and IV administration of TXA. The availability of pharmacokinetic comparisons in an animal model is the first step in developing evidence-based protocols.

Competing Interests

The authors declare that they have no conflicts of interest.

Acknowledgments

Cheryl Meek, Lindsay Burrowes, John Tyberg, Kendal Weatherby, Anukul Ghimire.

References

AJEM