• Open Access

Poor Reproducibility of Template Bleeding Time in Horses


  • D. Segura,

    1. Servei de Medicina Interna Equina, Departament de Medicina i Cirurgia Animals, Facultat de Veterinària, Universitat Autònoma de Barcelona, Barcelona, Spain
    Search for more papers by this author
  • L. Monreal

    1. Servei de Medicina Interna Equina, Departament de Medicina i Cirurgia Animals, Facultat de Veterinària, Universitat Autònoma de Barcelona, Barcelona, Spain
    Search for more papers by this author

Corresponding author: Luis Monreal, Servei de Medicina Interna Equina, Departament de Medicina i Cirurgia Animals, Facultat de Veterinària, Universitat Autònoma de Barcelona, 08193-Bellaterra, Barcelona, Spain; e-mail: lluis.monreal@uab.es


Background: Template bleeding time (TBT) is considered to be a useful test for detecting platelet function disorders and the effect of platelet-activating drugs, but studies in human medicine have concluded that the test has poor reproducibility and sensitivity.

Hypothesis: TBT has poor reproducibility in horses and has insufficient sensitivity to detect the effect of etamsylate on platelet function.

Animals: Twenty healthy horses.

Methods: TBT was determined and repeated 2 hours and 30 days later. TBT was also performed 2 hours after IV administration of etamsylate.

Results: Although no statistical differences were seen between the TBT values obtained at different times, the coefficients of variation for TBT replicates ranged from 26.8% to 45.5%. The reference range for TBT was 138.4–860.4 seconds. No statistically significant shortening of the mean TBT value was observed after etamsylate administration.

Conclusion and Clinical Importance: TBT has poor reproducibility, and the reference range is too wide to make TBT useful in a clinical setting. Other tests with higher reproducibility should be considered when assessing platelet function disorders in horses.

Template bleeding time (TBT) was the first test used for platelet function assessment, and it is easy and quick to perform. It has the advantage of assessing primary hemostasis in vivo, including the role of the vessel wall in the hemostatic process, and involves no processing of blood.1

In humans, TBT is markedly prolonged in patients with von Willebrand disease and platelet function defects,2 as well as after the administration of some antiaggregant drugs like acetylsalicylic acid (ASA).3 On the other hand, administration of proaggregant drugs like etamsylate produces a significant shortening of TBT in humans, rabbits, pigs, and cows.4–6,a However, TBT progressively increases as the platelet count decreases, and it is not a useful test in thrombocytopenic patients.1 In addition, it has been observed that the sensitivity and specificity of the test to predict bleeding risks in humans are low, the reproducibility is poor, and TBT may be affected by technician skills and the skin thickness or temperature.1,7 Consequently, TBT is currently being replaced in human hospitals by new accurate tests, such as a platelet function analyzer (PFA-100)b.

In horses, the technique has been standardized and has been used to test the influence of different drugs on platelet function inhibition.8 Different studies have demonstrated that TBT is prolonged after ASA administration in horses,8–12 although no changes in TBT have been observed after treatment with other drugs such as flunixin meglumine or phenylbutazone,8 or after administration of unfractioned and low-molecular-weight heparins.13

Despite the fact that TBT has been used in horses in different clinical situations, and that it is considered to be a reproducible test by some authors,8 there are no reproducibility studies or reference intervals available, and the mean TBT values differ greatly among healthy animals in different studies. Because TBT has been shown to have poor reproducibility in humans,7 we hypothesized that TBT was not reproducible in horses and would not be able to detect the effect of etamsylate administration in this species. Consequently, the objective of this study was to evaluate whether TBT in horses has an acceptable reproducibility and to compare the range of TBT values obtained in our study with those from previous studies. Additionally, because TBT has been shown to be shortened after etamsylate administration in humans and several animal species, the effect of etamsylate on TBT was also evaluated.

Materials and Methods


Twenty healthy male horses of different breeds and ages (age range, 2–15 years) undergoing a daily exercise program were included in the study. Horses were considered to be healthy because they did not have a previous history of hemostatic disorder, and their physical examination, hemogram (CBC, PCV, total plasma protein, and fibrinogen concentrations), prothrombin time (PT), and activated partial thromboplastin time (aPTT) were within the reference ranges. Horses had been stabled under the same conditions for at least 1 year and had not received any pharmacological treatment for at least 4 weeks before the study. All procedures were conducted in compliance with the guidelines established by the Autonomous University of Barcelona Human and Animal Experimentation Ethics Committee.

Study Design

The day before TBT determination, a 4 × 4 cm area of skin was shaved medially in the middle portion of both forelimbs, a few centimeters proximal to the chestnuts (torus carpeus).

TBT was first determined in the right forearm (TBT-0). Two hours later, TBT was repeated (TBT-2), in the left forelimb. A third TBT determination was performed 30 days later (TBT-30) to allow healing of the wounds generated by the previous determinations. Immediately after, all horses received 12.5 mg/kg BW of etamsylate IV. TBT was repeated 2 hours later using the contralateral forelimb. In all cases, TBT determinations were performed by the same investigator (LM).

TBT determination was performed following the technique described by Kopp et al.8 Briefly, standard venostasis was achieved by placing a blood pressure cuff proximal to the sampling area and inflating it to a pressure of 40 mmHg for 1 minute. Immediately after, TBT was performed in duplicate by simultaneously using 2 automated devicesc, which produced 4 standardized cuts of 5 mm length and 1 mm depth in the skin, avoiding visible superficial blood vessels. The wounds were gently blotted every 30 seconds avoiding contact with the wound edges, and the time taken for bleeding to stop was recorded for each of the 4 cuts. The scars of previous cuts were avoided in subsequent determinations. TBT for each individual horse was determined as the mean of the time taken for bleeding to stop obtained for each of the 4 cuts.

Blood Samples

Blood samples were obtained immediately after each TBT determination to ensure that the bleeding times were not influenced by changes in platelet count or clotting defects.

Samples were obtained via a jugular venipuncture and placed in 5 mL evacuated tubes containing 3.8% sodium citrate for assessing clotting times and EDTA K3 for platelet count. EDTA samples were analyzed the same morning of sampling. Citrated samples were centrifuged and plasma was frozen at −20 °C until analysis. The time elapsed between sampling and processing was always <4 hours.

Platelet counts were determined by a semi-automated cell counter.d PT and aPTT were determined in duplicate using a semiautomatic coagulometere with commercial reagents, and the determination was repeated whenever the coefficient of variation between replicates was >5%.

Data Analysis

Comparison of Means

Normal distribution of data was tested by the Kolmogorov–Smirnov test. An ANOVA for repeated measures was used to compare TBT mean values at different times (TBT-0, TBT-2, and TBT-30). The mean values of clotting times and platelet count at the same times were compared using the same analysis.

The mean TBT values before and after etamsylate administration were compared by Student's t-test for paired samples; the mean PT, aPTT, and platelet count values before and after etamsylate administration were also compared using the same test. All results are presented as mean ± SD. Statistical significance was accepted for P <.05.

Overall Reproducibility

Between-replicate coefficients of variation (CV) were calculated at TBT-0 and TBT-2; TBT-0 and TBT-30; TBT-2 and TBT-30; and TBT-0, TBT-2, and TBT-30 using the following formula:

Between-replicate CV=


where SSQ denotes the between-replicate square sum, u the number of samples (animals), and n the number of replicates.


All horses tolerated the TBT procedure well, including the blood pressure cuff placement and the incising devices' discharge, and overt signs of anxiety were not observed in any horse.

The results for TBT-0, TBT-2, TBT-30, and TBT after etamsylate are shown in Figure 1. All series of data were normally distributed. The anova for repeated measures did not show statistically significant differences among TBT at the different time points (P=.4). However, the between-replicate CV ranged from 26.8% to 45.5% depending on the time points being compared.

Figure 1.

 Box plots of template bleeding time in seconds at time 0, repeated after 2 hours and 30 days, and after the administration of etamsylate. Center line is the median; interquartile ranges and 95% range are represented by the box and whiskers, respectively.

Because no statistically significant differences in TBT means were detected at different time points (0, 2, and 30), repeated measures from each horse were averaged and the range of TBT for all horses was calculated from these average values for comparison with other studies (see Table 1).

Table 1.   Normal values of template bleeding time (TBT) in healthy horses obtained from different studies.
Authors (year)TechniquenMeanSDMinMax
  • References are ranked according to the publication year. All available data were used. Missing information was not found in the publications. n, number of horses; Min, minimum; Max, maximum.

  • *

    Approximate TBT mean value taken from graphics.

  • The values obtained in this study are also shown.

Judson and Barton (1981)Duke4175.8
Trujillo et al (1981)*Duke10120
Kopp et al (1985)Kopp9468150840
Cambridge et al (1991)Kopp3648103.4350855
Monreal et al (1995)Kopp1026281.6148.5435.5
Taylor et al (2000)Kopp6212.451.6
Present studyKopp20 (× 3)341.7113.9136.3893.8

No statistically significant differences were found when comparisons were made between TBT mean values before and after administration of etamsylate (Fig 1).

Platelet counts and clotting times always remained within reference ranges. No statistically significant differences were found in PT, aPTT, and platelet count means at the different time points or before and after etamsylate administration (Table 2).

Table 2.   Mean ± SD values of platelet count (Plt) and clotting times (PT and aPTT) determined at the different time points (TBT-0, TBT-2, and TBT-30) and 2 hours after the intravenous administration of 12.5 mg/kg BWT of etamsylate in 20 healthy horses.
 TBT-0TBT-2TBT-30TBT after etamsylate
  1. Reference ranges: Plt, 100–300 × 103/μL; PT, 9–13 seconds; aPTT, 25–45 seconds.

  2. aPTT, activated partial thromboplastin time; PT, prothrombin time; TBT, template bleeding time.

Plt (× 103/μL)120.6 ± 37.0132.0 ± 23.2130.1 ± 31.6125.5 ± 24.6
PT (seconds)10.5 ± 0.410.4 ± 0.410.4 ± 0.810.2 ± 0.6
aPTT (seconds)30.9 ± 5.033.1 ± 5.031.4 ± 6.132.7 ± 4.0


Some authors have described TBT as a reproducible in vivo test to assess platelet function in horses.8 Nevertheless, when comparing TBT values in healthy horses reported in different studies (Table 1), it is not easy to decide which of the reported ranges can be considered reliable because of the high variability in TBT values. In addition, in the few studies available, data are often incomplete and very few horses were used in most of them. Data from Judson and Barton9 and Trujillo et al10 have been included in the table for completeness, but they cannot be compared with TBT values from other studies because of the different technique used in bleeding time determination.

The results of the present study regarding TBT are similar in terms of minimum and maximum to those obtained by Kopp et al,8 although the mean TBT value reported by these authors was approximately 126 seconds longer than the average TBT value observed in our study. Cambridge et al11 reported a mean TBT value in control horses that was still longer than the one reported in the present study, whereas Monreal et al13 obtained a mean TBT value <300 seconds in healthy control animals in another study, which was similar to that reported by Taylor et al.12 Considering that a similar technique was used to determine TBT in all these studies, the differences in the mean TBT values (from 212 to 648 seconds) do not seem to indicate good reproducibility.

Although TBT was determined in duplicate in each horse, the bleeding time was averaged from 4 standardized cuts, and the procedure was repeated at different time points, more animals should have been included to establish a reference range according to international guidelines.

To our knowledge, there are no previous articles reporting CV or reproducibility for TBT determination in horses. In our study, the overall reproducibility of TBT determination was unacceptable, even considering the biological complexity of the measurement. Although no significant differences were obtained when TBT mean values were compared at different time points, most of the horses showed a high variability in TBT even when it was performed only 2 hours apart by the same investigator.

Therefore, the interpretation of a TBT value in a clinical setting in which a control group is not available could be very difficult, and, on the other hand, it would be impractical to test several healthy horses each time TBT needs to be determined in a patient. In addition, the reported SD values are very high, which implies that large variations in TBT would be necessary before they could be detected by the technique. Nevertheless, TBT has been used successfully in horses to detect the effects of ASA administration on hemostasis. Prolongations between 50% and more than 100% of the basal TBT value have been observed after ASA administration at different dosing regimes.8–12

Etamsylate is a prohemostatic drug that has been demonstrated to produce moderate platelet activation in horses14 and humans15 and a significant shortening of TBT in human, experimental animals, and cattle.4–6,a However, no statistically significant differences were observed in our horses on comparing mean TBT before and after etamsylate administration. Because of the high SD values obtained, a difference of 115 seconds in the mean TBT value would have been required to detect any effect of etamsylate on TBT in this study with a statistical power of 0.8, which would probably have implied a dangerous thrombotic situation. Moreover, even such a marked difference would still have been within the range obtained in our study for healthy animals. Consequently, other methods should be recommended to assess platelet function and alterations in primary hemostasis in horses. For instance, a PFA-100 has been evaluated in horses,16 showing a much better reproducibility (11.8%) and lower SD values. This analyzer is considered to be the substitute for TBT in human medicine.17 However, other techniques such as flow cytometry have also been demonstrated to be much more accurate to assess equine platelet activation,18 and new platelet activation markers are under investigation in horses.19

In conclusion, the results of our study show that TBT has poor reproducibility in horses because of high variations among individuals and among determinations in the same individual, even when platelet counts remain unaltered. Moreover, the wide range of values observed in healthy animals limits the usefulness of TBT in clinical settings and research studies, unless marked prolongations are expected due to the effect of treatment or disease under investigation.


aHomedes J. El etamsilato como fármaco hemostático en la clínica del bovino. Barcelona: 2002. PhD Thesis

bPFA-100, Dade Behring Inc, West Sacramento, CA

cSimplate II, Organon Teknika Corp, Durham, NC

dSysmex F-800, TOA instruments, Japan

eStago ST4, Stago Diagnostics, Asnières-Sur-Seine, France


This study was financially supported by Laboratorios Dr Esteve SA.