Effects of 6% hetastarch (600/0.75) or lactated Ringer’s solution on hemostatic variables and clinical bleeding in healthy dogs anesthetized for orthopedic surgery


Amandeep Chohan, Department of Veterinary Clinical Sciences, College of Veterinary Medicine, PO Box 646610, Washington State University, Pullman, WA 99164-6610, USA. E-mail: chohan@vetmed.wsu.edu


Objective  To evaluate and compare hemostatic variables and clinical bleeding following the administration of 6% hetastarch (600/0.75) or lactated Ringer’s solution (LRS) to dogs anesthetized for orthopedic surgery.

Study design  Randomized blinded prospective study.

Animals  Fourteen, healthy adult mixed-breed hound dogs of either sex, aged 11–13 months, and weighing 20.8 ± 1.2 kg.

Methods  The dogs were randomly assigned to receive a 10 mL kg−1 intravenous (IV) bolus of either 6% hetastarch (600/0.75) or LRS over 20 minutes followed by a maintenance infusion of LRS (10 mL kg−1 hour−1) during anesthesia. Before (Baseline) and at 1 and 24 hours after bolus administration, packed cell volume (PCV), total protein concentration (TP), prothrombin time (PT), activated partial thromboplastin time (APTT), von Willebrand’s factor antigen concentration (vWF:Ag), factor VIII coagulant activity (F VIII:C), platelet count, platelet aggregation, colloid osmotic pressure (COP) and buccal mucosal bleeding time (BMBT) were measured. In addition a surgeon who was blinded to the treatments assessed bleeding from the incision site during the procedure and at 1 and 24 hours after the bolus administration.

Results  Following hetastarch or LRS administration, the PCV and TP decreased significantly 1-hour post-infusion. APTT did not change significantly compared to baseline in either treatment group, but the PT was significantly longer at 1-hour post-infusion than at 24 hours in both groups. No significant change was detected for vWF:Ag, FVIII:C, platelet aggregation or clinical bleeding in either group. The BMBT increased while platelet count decreased significantly at 1-hour post-infusion in both groups. The COP decreased significantly in both treatment groups 1-hour post-infusion but was significantly higher 1-hour post-infusion in the hetastarch group compared to the LRS group.

Conclusions and clinical relevance  At the doses administered, both hetastarch and LRS can alter hemostatic variables in healthy dogs. However, in these dogs undergoing orthopedic surgery, neither fluid was associated with increased clinical bleeding.


Crystalloid fluids like lactated Ringer’s solution (LRS) and synthetic colloids like hetastarch provide rapid volume expansion and are commonly used to maintain or restore intravascular volume in patients undergoing general anesthesia. Hetastarch can provide very effective and long lasting plasma volume expansion as well as colloidal support but can impair hemostasis (Kozek-Langenecker 2005; Chan 2008). Crystalloids can also cause dilution of clotting factors (Langer et al. 1998; Sartain 2003) but have also been shown to cause hypercoagulability (Kirlay et al. 2006). However, the alterations caused by hetastarch have been shown to occur not just via dilution effects but also due to direct effects on coagulation factors and platelet function (Stump et al. 1985; Jamnicki et al. 2000; Franz et al. 2001; Strauss et al. 2002; Smart et al. 2009).

Hetastarch is a type of hydroxyethyl starch (HES) solution manufactured by hydroxyethylation of glucose molecules at carbon positions 2, 3 and 6 of the branched natural polymer, amylopectin. Hydroxyethylation stabilizes the polymer, increases solubility and slows down the plasma degradation by interfering with amylase activity (Treib et al. 1999; Evert de Jonge & Levi 2001). Hetastarch has a high degree of substitution, that is, a high proportion of glucose molecules with substituted hydroxyethyl groups (between 0.6 and 0.75) and high in vitro molecular weight (450,000–600,000 kDa), and has been shown to impair coagulation to a greater extent than other HES preparations such as pentastarches (0.5) and tetrastarches (0.4) (Strauss et al. 2002; Jungheinrich et al. 2004; Gandhi et al. 2005). Yet, hetastarch is the most common preparation of HES available and is widely used for volume expansion and/or colloidal support in North America (Treib et al. 1999; Wierenga et al. 2007; Chan 2008).

The guidelines for use of hetastarch in veterinary medicine have been extrapolated from human safety guidelines rather than from veterinary specific studies. In the veterinary literature, there is a scarcity of studies evaluating the effect of hetastarch on hemostasis in vivo in anesthetized dogs. A recent in vivo study in non-anesthetized dogs showed that a 20 mL kg−1 dose of hetastarch (670/0.75) could prolong platelet closure time for up to 24 hours (Smart et al. 2009). Another previous study in healthy anesthetized dogs revealed that use of 6% hetastarch at 6 mL kg−1 for resuscitation after induced hypovolemia was not associated with any clinical bleeding (Zoran et al. 1992). The present study was performed to compare the effects of a bolus (10 mL kg−1, IV) of either 6% hetastarch (600/0.75) followed by LRS infusion or LRS bolus (10 mL kg−1, IV) followed by LRS infusion at a rate of 10 mL kg−1 hour−1 on various laboratory hemostatic variables and clinical bleeding in dogs undergoing general anesthesia for stifle arthroscopy. We hypothesized that the effects on coagulation caused by 6% hetastarch (600/0.75) will be more pronounced in comparison to LRS.

Materials and methods

Fourteen healthy, adult, purpose-bred hound dogs of either sex between 11 and 13 months of age were studied. Within 1 month of the study, the dogs had not received any medication known to alter coagulation tests or other variables being measured. All dogs were fed commercial dog food and received water ad libitum. The Washington State University Institutional Animal Care and Use Committee approved all animal protocols (ASAF # 3777).

Experimental protocol

On the day of the study, food, but not water, was withheld from the dogs for 12 hours. Approximately 1 hour prior to the surgery, a thorough physical examination was performed and baseline data were collected (see sampling section below). Dogs were then premedicated with hydromorphone (0.1 mg kg−1, IM). Twenty to 30 minutes after premedication, a 20-gauge 1.25-inch (2.5 cm) IV catheter was placed in the cephalic vein for administration of anesthetic induction drugs and either 6% hetastarch (Hespan; Braun Medical Inc, CA, USA) having 0.75 degree of substitution and average molecular weight of 600 kDa or LRS (Hospira Inc, IL, USA). Induction of anesthesia was achieved using thiopental sodium (8.6 ± 1.4 mg kg−1) administered IV to effect. The dogs were endotracheally intubated and maintained on isoflurane in oxygen delivered to effect via a standard small animal anesthetic machine and circle breathing system. The L7-S1 area was clipped of hair and prepared for aseptic lumbosacral epidural injection using preservative free morphine (0.1 mg kg−1) diluted with sterile saline to a total volume of 0.22 mL kg−1 body weight. All dogs underwent either left or right stifle arthroscopy as part of a separate study.

The study was originally designed for 16 dogs randomly assigned by a coin toss to either of the treatment groups in order to have eight dogs in each group. However, the final two dogs to be assigned became unavailable late in the course of the study resulting in unequal distribution of dogs in the two treatment groups. Six dogs received an LRS bolus (10 mL kg−1) and eight dogs received a 6% hetastarch bolus (10 mL kg−1). Bolus administration was completed in approximately 20 minutes in both groups. All dogs then received an LRS infusion at a rate of 10 mL kg−1 hour−1 for the duration of the surgery. The bolus was administered while the dogs were clipped and prepared for the epidural and stifle arthroscopic surgery. Average duration of time from the start of the bolus to the time when the 1-hour post-infusion sample was collected in the hetastarch group was 80 minutes while it was 78 minutes in the LRS group. All the procedures were approximately 1-hour long and the dogs were still under anesthesia when the 1-hour post-infusion sample was collected.

Blood sample collection

Approximately 1 hour prior to surgery the dogs were physically restrained without sedation and buccal mucosal bleeding time (BMBT) was measured using a spring-loaded cutting blade (Simplate II; Organon Teknika Co, NC, USA). A 20-gauge, 1.5-inch (3.8 cm) needle was inserted into the jugular vein and approximately 9 mL of blood was withdrawn into a non-heparinized 12 mL polypropylene syringe for analysis of hemostatic variables including packed cell volume (PCV), total protein (TP), platelet count, prothrombin time (PT), activated partial thromboplastin time (APTT), von Willebrand factor antigen concentration (vWF:Ag), factor VIII coagulant activity (FVIII:C), platelet aggregation, and colloid osmotic pressure (COP). Blood was also collected at 1 and 24 hours by needle aspirate from the jugular vein. The blood was immediately divided into four tubes: 1 mL into an EDTA-containing tube (for platelet determination), 5.4 mL divided equally into two 3 mL citrate tubes containing 0.3 mL of 3.8% trisodium citrate (for determination of PT, APTT, vWF:Ag, F VIII:C and platelet function studies) and 2 mL into a heparin vial (for COP determination). For determination of PCV and TP, blood was directly collected in two microcapillary tubes at each sampling time. The PCV was determined by microhematocrit procedure and TP concentration was measured by a refractometer (Schuco Clinical Refractometer Model S2020; Allied Healthcare Products Inc, MO, USA) Blood samples for the determination of COP, vWF:Ag concentration and F VIII:C activity were immediately centrifuged to separate the plasma, rapidly frozen, and stored at −70 °C.

Coagulation procedures

The coagulation times (PT, APTT) were determined on all samples by use of a semi-automated hemostasis analyzer (Model Start 4; Diagnostica Stago, NJ, USA) based on the method of electro-mechanical clot detection (viscosity based detection system). Both APTT and PT analysis were performed according to the manufacturer’s recommendations. PT was determined with rabbit brain thromboplastin from a commercial source (Simplastin Excel; Trinity Biotech PLC, Ireland). Commercially available activated partial thromboplastin reagent (Rabbit brain cephalin; Dade Actin Activated Cephaloplastin Reagent, Dade Behring, Siemens Healthcare Inc., Germany) was used to evaluate APTT. All the samples were analyzed within 2 hours of collection. The vWF:Ag concentration was determined by using an ELISA, configured with monoclonal anti-canine vWF antibodies (Comparative Hematology Section, Diagnostic Laboratory, College of Veterinary Medicine, Cornell University, Ithaca, NY, USA). The FVIII:C activity assay was performed using a modified one-stage APTT technique with a commercial APTT reagent (Dade Actin; Baxter Diagnostics, NJ, USA) and canine congenital factor VIII deficient substrate plasma. Clotting times were determined for test plasmas and reported as percentage activity after log-log transformation based on the value of a canine plasma standard having an assigned value of 100% FVIII:C activity (Comparative Hematology Section, Diagnostic Laboratory, College of Veterinary Medicine, Cornell University, Ithaca, NY, USA). The inter-assay coefficient of variation is <10% for FVIII:C activity and <8% for vWF:Ag concentration (Benson et al. 1991; Brooks et al. 1996; Stokol et al. 2000; Wardrop & Brooks 2001).

Platelet aggregation determinations

Platelet aggregation, in response to platelet activating factor (PAF; Tocris Bioscience, MO, USA), was measured using an electronic aggregometer (Model 700; Chrono-Log Co, PA, USA). A 500-μL aliquot of citrated whole blood was mixed with 500 μL of normal saline and incubated for 5 minutes at 37 °C. After equilibration for 1 minute at 37 °C, 50 μL of 0.1 or 10 nm of PAF was added. Platelet response to 0.1 and 10 nm PAF was measured as a change in electrical impedance (in ohms) over 7 minutes. Platelet aggregation was performed for baseline and 1-hour post-infusion blood samples from four dogs in the hetastarch group and three in the LRS group within 1 hour of sample collection.

Buccal mucosal bleeding time determinations

The procedure was performed with the dogs in lateral recumbency with minimal restraint. The upper lip was everted and held in place. The cutting device was used to make a vertical incision of standard length and depth directly above the maxillary canine tooth in an area free of any obvious vasculature. Timing was begun as soon as the mucosal incision was made. Blood oozing from the incision was blotted using a circular filter paper held approximately 1–2 mm away from the incision, taking care not to disrupt the clot formation. The timing was continued until the bleeding ceased from the incision and the BMBT was recorded in seconds. One operator (ASC) performed all BMBTs to reduce operator error.

Colloid osmotic pressure determinations

Frozen plasma samples were thawed and COP was determined by using a colloid osmometer (Model 4420; Wescor Inc, UT, USA). The osmometer was set to zero with normal saline, and then calibrated with processed and stabilized bovine serum (Osmocoll; Wescor Inc) as reference standard. Both high (23.9–27.9 mmHg) and low (14.4–16.4 mmHg) controls were used for calibration. COP was determined for all the blood samples.

Clinical bleeding assessment

An experienced orthopedic surgeon who was blinded to the treatments assessed amount of bleeding during the procedure, 1-hour and 24 hours post-infusion of hetastarch or LRS bolus. Bleeding was described on a scale where 0 = no bleeding, 1 = amount of bleeding expected for the procedure, 2 = moderate bleeding, but excessive for the procedure and 3 = frank hemorrhage.

Statistical analysis

The MIXED procedure in sas was used to perform an anova test for the effects of time and treatment at an overall significance level of 0.05. Fisher’s LSD test was selected for post hoc analysis of these data. This parametric procedure is robust to violations of the assumptions of normality and was deemed appropriate for analysis of time and treatment interaction effects in this statistical model. Non-normally distributed data (arithmetic differences from baseline at 1-hour post-infusion for all variables) were analyzed using the Wilcoxon rank sum test.


The results are reported as the arithmetic mean ± standard deviation (range) except for non-normally distributed data that are expressed as median (range). There were no significant differences between groups for baseline hemostatic variables, patient age, weight, surgical time and total volume of fluids (colloid + crystalloid) administered. All baseline values were within the reference interval for our laboratory except for two dogs in the hetastarch group and three dogs in the LRS group, each of which had baseline PTs higher than the established reference intervals (6.4–8.2 seconds) for our laboratory (Table 1). Within each treatment group, both PCV and TP were significantly (p < 0.0001) lower at 1-hour post-infusion compared to both baseline and 24 hours post-infusion. The BMBT increased significantly (p = 0.0082) at 1-hour post-infusion in both groups compared to baseline (Table 1). No significant changes from baseline were noted at any time in either of the groups or in between groups for APTT, vWF:Ag concentration (Fig. 1) and FVIII:C activity (Fig. 2) while PT was significantly (p = 0.002) higher 1-hour post-infusion compared to 24 hours post-infusion in both groups.

Table 1.   Mean ± SD (range) for hemostatic variables in fourteen dogs prior to and at 1 and 24 hours post-infusion of a 20-minute IV administration of 10 mL kg−1 of either hetastarch or LRS, each followed by a maintenance infusion of LRS (10 mL kg−1 hour−1)
Hetastarch (n = 8)LRS (n = 6)
Baseline1-hour post-infusion24-hour post-infusionBaseline1-hour post-infusion24-hour post-infusion
  1. *Significantly different from baseline, different superscript letters show significant differences at that time.

  2. †Significantly different from value at 1-hour post-infusion.

TP (g dL−1)7.1 ± 0.3 (6.8–7.6)5.7 ± 0.3* (5.2–6.2)6.8 ± 0.5† (6.2–7.6)7.1 ± 0.3 (6.6–7.6)5.5 ± 0.4* (5.0–6.0)6.9 ± 0.5† (6.2–7.6)
PCV (%)51 ± 2 (48–55)38 ± 4* (30–45)51 ± 3†(46–55)53 ± 4 (50–60)41 ± 2* (38–43)51 ± 6† (44–60)
Platelets (103 μL−1)261 ± 37 (225–335)222 ± 43* (158–280)241 ± 92† (47–337)284 ± 100 (167–367)201 ± 100* (120–288)334 ± 68† (216–407)
BMBT (seconds)107 ± 20 (80–141)151 ± 46* (112–260)111 ± 22 (92–159)112 ± 24 (85–154)130 ± 25* (101–176)127 ± 21 (102–155)
PT (seconds)7.8 ± 0.8 (6.7–9.1)9.2 ± 2.8 (7.2–16.0)6.9 ± 0.5† (6.4–7.3)8.1 ± 1.1 (6.9–9.9)8.5 ± 0.9 (7.4–9.9)7.3 ± 0.3† (6.9–7.3)
APTT (seconds)11.2 ± 0.6 (10.4–11.9)12.7 ± 2.7 (10.9–19.1)11.5 ± 1.0 (10.2–13.2)11.8 ± 1.7 (9.7–14.0)11.7 ± 1.3 (10.2–14.0)11.1 ± 0.8 (9.9–12.0)
COP (mmHg)21.5 ± 1.2 (19.6–23.2)18.5 ± 1.4*a (15.7–20.2)21.4 ± 2.1† (17.4–23.6)20.6 ± 3.3 (14.6–23.2)15.8 ± 2.2*b (13.5–19.6)22.2 ± 2.4† (19.9–26.1)
Figure 1.

 Mean and SD von Willebrand factor concentration as a percent of normal pooled plasma samples in dogs after administration of 10 mL kg−1 bolus of either hetastarch or LRS over 20 minutes, each followed by a maintenance infusion of LRS (10 mL kg−1 hour−1) in healthy dogs undergoing stifle arthroscopy.

Figure 2.

 Mean and SD factor VIII:C activity values as a percent of normal pooled plasma samples in dogs after administration of 10 mL kg−1 bolus of either hetastarch or LRS over 20 minutes, each followed by a maintenance infusion of LRS (10 mL kg−1 hour−1) in healthy dogs undergoing stifle arthroscopy.

The COP significantly (p = 0.0001) decreased from baseline at 1-hour post-infusion in both groups. At 1-hour post-infusion, COP was significantly (p = 0.0265) higher in the hetastarch group compared to the LRS group. At 24 hours post-infusion, COP was significantly higher than 1-hour post-infusion sample in both groups (Table 1).

Platelet count was significantly decreased at 1-hour post-infusion compared to baseline (p = 0.04) and 24 hours (p = 0.01) post-infusion in both groups. There was no significant change in platelet aggregation in either group or between groups (Table 2). All dogs scored a 1 on the clinical bleeding scale during the surgery and at 1 and 24 hours after the bolus of either LRS or hetastarch.

Table 2.   Platelet aggregation in response to 0.1 and 10 nm concentrations of platelet activating factor (PAF) as assessed by change in impedance in seven dogs prior to and at 1 and 24 hours post-infusion of a 20-minute IV administration of 10 mL kg−1 of either hetastarch or LRS, each followed by a maintenance infusion of LRS (10 mL kg−1 hour−1)
Platelet aggregationGroup
Hetastarch (n = 4)LRS (n = 3)
Baseline1-hour post-infusionBaseline1-hour post-infusion
PAF conc.0.1 nm10 nm0.1 nm10 nm0.1 nm10 nm0.1 nm10 nm
Impedance (Ω)

One of the planned comparisons in our study was to compare the two treatment groups with respect to the differences from baseline at 1-hour post-infusion. The data for the differences from baseline at 1-hour post-infusion were not normally distributed (Table 3). The Wilcoxon rank sum test did not show any statistically significant differences between the two treatment groups for this comparison in any of the measured variables.

Table 3.   Median (range) for arithmetic differences from baseline at 1-hour post-infusion (1-hour post-infusion value minus baseline value) for variables in dogs receiving 20-minute IV administration of 10 mL kg−1 of either hetastarch or LRS, each followed by a maintenance infusion of LRS (10 mL kg−1 hour−1)
PCV (%)−11 (−17 to −10)−13 (−21 to −7)
TP (g dL−1)−1.7 (−2.2 to −1.1)−1.4 (−2.2 to −0.8)
Platelet count (103 μL−1)−59.5 (−324 to 62)−47 (−82 to 18)
BMBT (seconds)16 (−19 to 82)45 (−12 to 152)
vWF:Ag concentration(%)−8 (−259 to 63)−34 (−242 to −8)
Platelet aggregation (impedance in Ω)−1(−2 to 1)1(0 to 6)
PT (seconds)0.8 (−1.9 to 1.8)0.7(−0.7 to 8.0)
APTT (seconds)0.2 (−3.6 to 2.7)0.4 (−0.6 to 7.9)
FVIII:C activity (%)−4 (−76 to 239)−30 (−62 to 541)
COP (mmHg)−5.7 (−7.6 to −0.8)−3.1 (−4.9 to −1.1)


In our study administration of both hetastarch and LRS was associated with increased BMBT but did not significantly decrease vWF:Ag concentration or FVIII:C activity in either of the groups. At 1-hour post-infusion COP was significantly higher in the hetastarch group compared to LRS group. In both treatment groups the effects from dilution were observed at 1-hour post-infusion.

The purpose of this study was to determine the effects of a 10 mL kg−1 hetastarch bolus on various hemostatic variables, COP and clinical bleeding in healthy dogs undergoing general anesthesia for an orthopedic procedure. This dose was chosen because it is a commonly recommended dose for both hetastarch and LRS in anesthetized patients (Gaynor et al. 1996; Kudnig & Mama 2002). Since all of the dogs received LRS regardless of which treatment group they were in, we recognize that our study is not a clean assessment of the effects of hetastarch on hemostasis. However, the purpose of our study was to evaluate the effects of hetastarch in a clinical setting. Most often, clinically anesthetized patients receive both hetastarch and LRS in response to hypotension, hypovolemia, and/or hypoalbuminemia. Thus, we chose to study the effects of a bolus of hetastarch followed by a standard infusion of LRS. The inclusion of a group receiving a bolus of LRS followed by a standard infusion of LRS was to allow comparison of the hetastarch group against a group experiencing primarily dilution effects from the administration of a fluid bolus.

In this study, a decrease in both PCV and TP occurred at 1-hour post-infusion in both groups and this could be attributed to hemodilution (Glowaski et al. 2003). The PCV and TP returned to baseline by 24 hours in both groups reflecting the resolution of dilution effects. In a previous study (Zoran et al. 1992) using hetastarch at 6 mL kg−1, dilution effects were not present at 24 hours post infusion. The duration of intravascular volume expansion for hetastarch has been shown to be from 12 to 48 hours in humans but beneficial effects in terms of volume expansion beyond 24 hours have not been proven (Kozek-Langenecker 2005). Compared to humans, normal dogs have an increased concentration of serum α-amylase, the enzyme responsible for plasma degradation of hetastarch (Yacobi et al. 1982), thus, hetastarch has a shorter half-life in dogs and, consequently, a shorter duration of effect in terms of volume expansion (Moore & Garvey 1996). Packed cell volume has been used to estimate blood volume (Glowaski et al. 2003). In the present study, the increase in blood volume, as estimated by PCV at 1-hour post-infusion in the hetastarch group, was 36% (51.4/37.4 = 1.36), whereas the LRS group had a 28% increase in blood volume from baseline value (52.5/40.8 = 1.28; see Table 1).

A refractometer calibrated for use with plasma was used to measure the total protein (TP) concentration in this study. The refractive index correlates well with the plasma protein concentration (Robini & Wolf 1956) and changes in refractive index parallel changes in plasma protein as shown by a good correlation between the TP measured by refractometer and TP measurements in clinical studies (Bowie & Owen 1982; Rackow et al. 1983). The relationship between colloid osmotic pressure and refractive index may be different for plasma protein versus synthetic colloid (Concannon 1992). There is a poor correlation between the changes in colloid osmotic pressure and the TP as measured by the refractometer after the administration of synthetic colloids like hetastarch and dextran (Thomas & Brown 1992; Bumpus et al. 1998). The TP as measured by refractometry is minimally affected and usually underestimates the COP values after synthetic colloid administration (Bumpus et al., 1998). Hence, a direct measurement of COP is advocated to measure the efficacy of colloid therapy.

There was a significant decrease in COP at 1-hour post-infusion in both the groups compared to baseline but the COP at this time point was significantly higher in the hetastarch group compared to the LRS group. COP decreased by 24% in the LRS group and 14% in the hetastarch group (Table 1). The decreased COP in the LRS group is most likely due to hemodilution. The COP of LRS has been shown to be 0 mm of Hg (Chan et al. 2001) so it just dilutes out the plasma proteins, hence causing the decrease in COP. Previous studies have shown either no change in COP (Muir & Wiese 2004) or a modest impact on COP (Smiley & Garvey 1994; Moore & Garvey 1996; Chan 2008) after administration of hetastarch. The COP of 6% hetastarch as used in this study is reported to be 32.7 ± 0.2 mmHg (Chan et al. 2001). The COP depends on the number of the particles. The seemingly paradoxical decrease in COP observed in our hetastarch group at 1 hour may be due to the administration of LRS after hetastarch administration that decreased the concentration of hetastarch molecules. During this hour some of the smaller hetastarch molecules were excreted through the kidneys thus further decreasing the number of particles. The relatively higher COP of the hetastarch would help to retain the LRS in the intravascular compartment. However, clearly dilution factors aren’t the only factors involved since volumes infused in each group were similar (386 ± 63 mL [bolus + infusion] for the hetastarch group and 397 ± 83 mL for the LRS group). The fact that the COP 1-hour post-infusion in the hetastarch group was higher than in the LRS group is likely due to the presence of large hetastarch molecules that contribute to higher COP. When a significant increase in COP occurs after hetastarch administration, the increase occurs rapidly after administration and dissipates quickly. One study in hypoalbuminemic dogs reported that COP had returned to pre-infusion values 12 hours after hetastarch administration (Moore & Garvey 1996). At 24 hours post-infusion there was no difference in COP either between groups or within a group compared to the baseline in our study. This lack of difference is possibly due to absence of dilution effects of the hetastarch 24 hours post-infusion, as the availability of osmotically active hetastarch molecules is reduced because of renal excretion of small molecular weight fractions and redistribution and metabolism of high molecular weight fractions.

Hemostasis is a complicated process involving dynamic interactions between vasculature, platelets and coagulation proteins (Troy 1988). The BMBT is probably the best measure of platelet plug formation and platelet function in vivo (Johnson et al. 1988). The BMBT is not reported to be prolonged until the platelet count decreases to less than 70,000 μL−1 (Stockham & Scott 2008). In addition, platelets require von Willebrand’s factor (vWF) to adhere to the subendothelial collagen, and this accounts for prolonged bleeding times in patients with von Willebrand’s disease (Jergens et al. 1987). Although statistically no significant difference was detected for vWF:Ag concentration in either of the groups at any time interval, the vWF concentration decreased by 27% in the LRS group, reflecting purely the dilution effects from LRS, and by 56% in hetastarch group which was beyond the expected dilution effects of 36% in this group. Previous studies have also indicated that the hetastarch induced decrease in vWF:Ag concentration could occur beyond the dilution effects (Langer et al. 1998; Kozek-Langenecker 2005; Van der Linden & Ickx 2006; Chan 2008). Large hetastarch molecules can form a complex with vWF and accelerated elimination of this complex has been considered as a pathogenic mechanism responsible for the adverse effects on factor VIII/vWF complex (Kozek-Langenecker 2005). In our study, the BMBT increased and platelet count was decreased significantly at 1-hour post-infusion in both groups. A likely explanation is dilution effects from both LRS and hetastarch that influenced both platelet count and vWF:Ag concentration. Also it should be kept in mind that animals can show increased bleeding tendencies even in the face of normal platelet count if the platelet dysfunction caused by hetastarch is significant. A previous study in dogs has also shown that BMBT increased and platelet count decreased after hetastarch administration (Zoran et al. 1992).

Platelet function was determined using the impedance method on whole blood. The impedance method is an accepted and practical way to reveal changes in platelet function on many samples from the same animal (Abbate et al. 1986). This technique allows assessment of platelet function in the presence of other blood constituents thereby more closely depicting the in vivo scenario of hemostasis (Thompson et al. 1983). Hetastarch preparations are thought to bind directly to the GP IIb/IIIa receptor on the platelet surface and inhibit the binding of fibrinogen to the receptor thus preventing outside to inside signaling, platelet up-regulation and ultimately, formation of the platelet plug (Wierenga et al. 2007). Hetastarch, having a high molecular weight and a high degree of molar substitution, is correlated with a greater decrease in platelet function compared to other HES solutions (Franz et al. 2001). In this study, platelet aggregation was not diminished in dogs administered either hetastarch or LRS. Two out of four dogs in the hetastarch group had a small increase in impedance; however this was not beyond the expected variation of this test. Impedance in one dog increased from 4 to 10 Ω. While the increase in impedance may represent a true increase in platelet aggregation, it is also possible that increased platelet aggregation was due to another factor, such as a difficult blood draw. The results of platelet aggregation in this study did not support decreased platelet function with hetastarch administration at a dose of 10 mL kg−1; however the low number of dogs in this part of the study precluded meaningful statistical analysis.

Closure times are an in vitro evaluation of platelet function in which the process of platelet adhesion and aggregation following a vascular injury is simulated (Mammen et al. 1998). In a previous in vivo study (Smart et al. 2009), 20 mL kg−1 hetastarch administration led to a significant increase in closure times at 3 and 5 hours post-infusion as measured by PFA-100 (Platelet Function analyzer-100; Dade Behring Inc, FL, USA) compared to the control saline (0.9%) group but the values returned to the reference interval 24 hours post-infusion. In an in vitro study in which a dose of 10 mL kg−1 hetastarch was simulated, no significant change in the closure times compared to baseline was observed (Wierenga et al. 2007). The method used to evaluate platelet function in our study (i.e. platelet aggregometry) is quite different from the one used in the previous studies (i.e. closure time measurements with PFA-100). Although a different methodology was used, if a dose of 20 mL kg−1 caused no change in platelet function (closure time) at 24-hour post-infusion then a dose of 10 mL kg−1 is also not expected to result in changes at that time. This is supported by the fact that BMBT, vWF:Ag concentration and platelet count returned to the baseline at 24 hours. In addition, the dogs in our study were under general anesthesia compared to the previous studies that were conducted in non-anesthetized dogs, which may have led to some unknown variability. We also had a small number of samples analyzed for platelet function and a larger number might be necessary to determine any significant effects of hetastarch.

In an in vivo study (Zoran et al. 1992) in dogs the BMBT and vWF concentrations began to improve 1 hour after discontinuing the infusion and returned to baseline 24 hours post-infusion. In our study also the BMBT, vWF concentration and platelet count had returned to baseline 24 hours post-infusion. Results from our study and from those of a previous in vivo study (Zoran et al. 1992) indicate that more significant effects on primary hemostasis could be present within the first hour post-infusion of hetastarch when the dose is 10 mL kg−1 or under. The infusion rate used in Zoran’s study was 6 mL kg−1 which is less than our infusion of 10 mL kg−1 but what makes this comparison logical is that in both the studies no clinical evidence of impaired primary hemostasis (e.g. petechiation, ecchymosis, hematoma or increased incisional bleeding) was observed at any time interval in spite of the fact that statistically significant changes compared to baseline were observed within the first hour. So if the dose is 10 mL kg−1 or less it is unlikely to cause clinically significant primary hemostatic defects in healthy dogs but caution is advised in patients with preexisting coagulopathy where these subclinical alterations in coagulation profile could precipitate clinical bleeding. Likely the effects on hemostasis could be more prolonged if the hetastarch dose is higher.

The PT evaluates the status of coagulation factors in the extrinsic (factor VII and factor III) and common (factors X, V, II, and I) coagulation pathways of secondary hemostasis. Two dogs in the hetastarch group and three dogs in the LRS group had higher baseline PT (reference interval 6.4–8.2 seconds). The reference intervals have been set for 90% of the population and 5% variability is expected on either side. Three out of these five dogs were within 5% of the higher reference limit that is quoted in the machine manual to be acceptable. The values in one of the other two dogs returned to 5% of the higher reference limit and within normal for the other dog at 1-hour post-infusion. There were no statistically significant differences between baseline and 1-hour post-infusion samples in either of the groups. The PT was significantly (p = 0.002) higher at 1 compared to 24 hours post-infusion in both groups. This increase in the PT at the 1-hour post-infusion in both groups is likely attributed to dilution of clotting factors (Stump et al. 1985; Chan et al. 2002), and absence of dilution effects at 24 hours post-infusion allowed the values to return to baseline.

Factor VIII:C is routinely measured as percent of activity with normal values ranging from 50% to 200%. Our population values ranged from 50% to 428%. Variation between animals in terms of FVIII:C activity could have obscured significant differences between the sets of observations related to the three time points. We measured baseline levels of FVIII:C for each dog and then designated this value as 100% of activity for that patient. We then measured subsequent FVIII:C activity and recorded the values as the percentage of fluctuation from baseline in response to hetastarch and/or LRS administration. This helped to better delineate the individual variation in activity with respect to time and treatment effect. In contrast to a previous study (Kozek-Langenecker 2005) that showed a decrease in the FVIII:C activity beyond the dilution effects, our study showed no significant change in FVIII:C activity at any time interval in either group. One possible explanation is that FVIII:C is an acute phase protein (Stirling et al. 1998; Topper & Prasse 1998) whose concentrations can be influenced by many factors like ongoing trauma or inflammation. It is possible that surgical trauma could have led to an increase in the level of FVIII:C activity by release of more FVIII and hence precluded any significant decrease that might have occurred in FVIII:C activity due to dilution in either group or due to a direct decrease in activity associated with hetastarch.

The APTT assesses the intrinsic (factors XII, XI, IX, and VIII) and common coagulation pathways. There was no statistically significant change in APTT from baseline in either group at any sampling interval. However, one dog in the hetastarch group had an increased APTT (19.1 seconds; reference interval 8.4–14.8 seconds) 1-hour post-infusion. This dog also had increased PT (16 seconds; reference range 6.4–8.2 seconds) at the same time point. The clinical relevance of these increased values from this single dog is questionable since this dog’s values returned to the normal range at 24 hours post-infusion and no increase in clinical bleeding was observed. Decrease in FVIII:C activity has been shown to cause an increase in APTT after hetastarch infusion (Evert de Jonge & Levi 2001) and this appears to be dependent on the concentration of HES in blood (Jungheinrich & Neff 2005; Van der Linden & Ickx 2006). There are two ways by which hetastarch can affect hemostasis. It is quite possible that the dilution effects in the hetastarch group due to the combination of the tested hetastarch dose (10 mL kg−1) and subsequently administered maintenance LRS infusion (10 mL kg−1 hour−1) in the present study were not enough to cause significant reduction in FVIII:C activity. Further, the infusion of LRS after the hetastarch bolus might have decreased the plasma concentration of hetastarch to an extent that it did not have direct effects on FVIII:C activity. Also, surgical trauma could have led to an increase in FVIII:C activity as it is an acute phase reactant (Topper & Prasse 1998). It is unlikely that the hemodilution experienced by the LRS group could have led to any significant reduction in FVIII:C activity as dilution was less when compared to the HES group (28% vs 36%). Coagulation factors must be reduced to at least 30% of baseline in order for PT and APTT to be prolonged (Mischke et al. 2003). Since APTT and PT did not increase significantly from baseline at any point in this study, it is unlikely that these dogs would have a clinical coagulopathy that is attributable to altered intrinsic or extrinsic pathways following hetastarch or LRS administration at the dose administered.

Bleeding during the procedure and incisional bleeding at 1 and 24 hours post-infusion was not more than expected and was similar in all the dogs. Clinically apparent hemostatic defects such as ecchymosis, petechiation and hematoma were not observed in a previous study using hetastarch at 6 mL kg−1 (Zoran et al. 1992) but increased incisional bleeding as well as increased bleeding into the body cavities has been documented in dogs receiving large doses of hetastarch (>30 mL kg−1) (Thompson & Gadsen 1965; Garzon et al. 1967). A previous study in non-anesthetized hypoalbuminemic dogs showed no relationship between worsening of the coagulogram and the dose of hetastarch used and stated that hetastarch related changes in the coagulogram are minor and of little clinical significance (Smiley & Garvey 1994). Our results from healthy anesthetized dogs support these findings, although we would agree that having a surgeon visually evaluate blood loss is not the most accurate way to assess the amount of hemorrhage since this assessment could vary among surgeons. A more accurate assessment could have been to actually measure the volume of blood loss.

We studied healthy normotensive, normovolemic-anesthetized dogs undergoing orthopedic surgery. Caution is advised when interpreting these results in reference to clinical situations involving hypovolemia. Adverse effects on hemostasis caused by hydroxyethyl starch solutions are determined by their in vivo pharmacokinetic behavior (Treib et al. 1999). Variable plasma volume expansion based upon preexisting volume status has been described (Metcalf et al. 1970). Thus, altered volume status and fluid dynamics in hypovolemic patients may result in variable plasma volume expansion and hence, the concentration of hydroxyethyl starch resulting in hemostatic side effects that differ from normal healthy animals.

We conclude that both LRS and hetastarch can cause changes in markers of hemostasis in healthy dogs. However, the results of this study suggest that, in these normal dogs anesthetized for stifle arthroscopy, the non-specific dilution effects of both 6% hetastarch and LRS on PT, platelet count and BMBT were not associated with increased clinical bleeding. Although not statistically significant, the specific effects of hetarstarch on vWF:Ag concentration and FVIII:C activity may have contributed to observed changes in measured hemostatic variables and caution is advised in patients who might have a pre-existing coagulopathy where hetastarch could potentiate a bleeding problem.