Assessment of hemostatic changes after crystalloid and colloid fluid preloading in trauma patients using standard coagulation parameters and thromboelastography.

BACKGROUND: The choice of an ideal fluid administered post trauma and its subsequent influence on coagulation still poses a clinical dilemma. Hence, this study was designed to assess the influence of in vivo hemodilution with various fluid preparations (4% gelatin, 6% hydoxyethyl starch (HES), Ringer's lactate, 0.9% normal saline) on coagulation using standard coagulation parameters and real-time thromboelastography (TEG) in patients undergoing elective surgery post trauma.
METHODS: In a randomized, double-blind study, 100 patients of either sex and age, belonging to ASA Grades I and II, scheduled for elective surgeries were allocated into four groups of 25 each according to the type of fluid infused. Group G (4% gelatin), Group N (0.9% normal saline), Group R (Ringer's lactate), and Group H (6% HES) received preloading with 1 L of fluid according to the group. The coagulation status of the patients was assessed during perioperative period (before surgery, after fluid preloading, and at the end of the surgery) using both conventional coagulation analysis and TEG. STATISTICAL ANALYSIS: Analysis of variance (ANOVA), post hoc and Pearson Chi-square test were used.
RESULTS: In all the patients preloaded with gelatin, there was a significant increase in prothrombin time index (PTI; 14.88±0.90 vs. 13.78±3.01, P<0.001) and international normalized ratio (INR; 1.12±0.09 vs. 1.09±0.19, P<0.05) compared to the baseline value. An increase was observed in these parameters in the postoperative period also. In the HES group, there was statistically significant increase in PT time (15.70±1.51 vs. 13.74±0.75, P=0.01) and INR (1.20±0.15 vs. 1.03±0.17, P<0.001) as compared to the baseline. In the intergroup comparisons, the patients preloaded with HES had a significant increase in INR (1.20±0.15 vs. 1.12±0.09, P=0.04) and reaction time (R time; 6.84±2.55 min vs. 4.79±1.77 min, P=0.02) as compared to the gelatin group. The fall in coagulation time (k time; 2.16±0.98 vs. 3.94±2.6, P=0.02), rise in maximum amplitude (MA; 61.94±14.08 vs. 50.11±14.10, P=0.04), and rise in A20 (56.17±14.66 vs. 43.11±14.24, P=0.05) were more in patients preloaded with RL as compared to the HES group. 100% patients in the gelatin group, 84.2% patients in the NS group, 94.4% patients in the RL group, and 66.7% patients in the HES group had hypocoagulable (R time > 14 min) state in the postoperative period.
CONCLUSION: Crystalloids are optimal volume expanders in trauma, with RL having beneficial effects on coagulation system (decrease in k time and increase in MA and A20). Among the colloids, HES 6% (130/0.4) affects coagulation parameters (increase in PTI, INR, R time, k time) more than gelatin. Trial registration (protocol number-IEC/NP-189/2011).

RCT Entities:   Population Interventions Outcomes

INTRODUCTION

Trauma patients are known to develop dilutional coagulopathy attributable to blood loss, consumption of coagulation factors and platelets, and intravascular volume replacement.[1] Incidence of coagulopathy in trauma patients on admission is 25-35%.[2] The choice of an ideal fluid administered post trauma and its subsequent influence on coagulation still poses a clinical dilemma. Hence, this study was designed to assess the influence of in vivo hemodilution with various fluid preparations (4% gelatin, 6% hydoxyethyl starch (HES), Ringer's lactate (RL), 0.9% normal saline (NS)) on coagulation using standard coagulation parameters and real-time thromboelastography (TEG) in patients undergoing elective surgery post trauma. In view of the large number of retracted papers by Professor Boldt,[3] there is a need for renewed studies of the effects of HES solutions on coagulation, which is otherwise well-traveled ground.

METHODS

Ethical approval for this study (protocol number IEC/NP-189/2011) was provided by the Ethical Committee of AIIMS (Chairperson Prof. J. P. Wali) on 8 August 2011. Written and informed consent was taken from the patients prior to surgery. The study was a prospective randomized, comparative, double-blind study conducted in 100 trauma patients above 20 years of age, posted for elective surgery in a referral tertiary trauma center. Patients on anticoagulants or antiplatelets were excluded from the study. All the blood samples which were not in proper proportions to the anticoagulant, hemolyzed samples, or samples collected by venipuncture taking more than 30 seconds were also excluded.

Study groups

Patients were randomly divided based on computer-generated randomization list into four groups according to the type of fluid infused. Group G (4% gelatin), Group N (0.9% NS), Group R (RL), and Group H (6% HES) received preloading with 1 L of fluid within 45 min according to the group after induction of anesthesia. Anesthesiologist preloading the fluid was also blinded to the type of fluid being infused. The following fluids were investigated:

4% succinylated gelatin solution, Gelofusine®, B. Braun Co., Penang, Malaysia

0.9% normal saline, Viaflex, Baxter, Gurgaon, India

Ringer's lactate, Viaflex, Baxter, Gurgaon, India

6% HES 130/0.4, Voluven, Fresenius kabi, Germany.

The coagulation status of the patients was assessed during perioperative period (before surgery, after fluid preloading, and at the end of the surgery in the recovery room) using both conventional coagulation analysis and TEG. Conventional coagulation parameters measured included prothrombin time index (PTI), activated prothrombin time (APTT), and international normalized ratio (INR). TEG parameters being measured were reaction time (R time), coagulation time (k time), alpha angle (α angle), maximum amplitude (MA), A10, and A20. The baseline characteristics (age, sex, diagnosis, mode of injury, type of anesthesia, Injury Severity Score (ISS), hemogram, and intraoperative blood loss) of all trauma patients undergoing elective surgery were recorded.

4.5 ml blood sample was collected by venipuncture into vacutainer tubes containing citrate (0.129 M trisodium citrate) for standard coagulation parameters and thromboelastogram, 2 ml collected in ethylenediaminetetraacetic acid (EDTA) vacutainer tubes for hemogram and platelet counts, and 0.36 ml (360 μl) whole blood was taken in automated thromboelastometer (TEM-A vacutainer). Blood samples were drawn before surgery, after fluid preloading, and at the end of the surgery. For PT, APTT and INR, the fully automated coagulation analyzer STA-COMPACT and STA reagents were used. R time, k time, α angle, MA, A10, and A20 were obtained using TEM-A automated thromboelastometer (Framar Biomedica, Rome, Italy) within 4 min of venipuncture. Analysis was performed in a standardized way and assessed for R time (normal range 9-14 min), k time (normal range 4-6.5 min), α angle (normal range 29°-43°), and MA (normal range 48-60 mm). Automatic calibration before every test and calibration for the end scale of 100 ensured the same linearity of the signal throughout its range, i.e., in TEM-A, it is sufficient to run a single quality control test that verifies the zero.

To study the correlation between TEG parameters and standard coagulation assays, we pooled together all the samples taken for TEM-A analyzer and compared these with standard coagulation assays.

Statistical analysis

A sample size of 60 is required to be within 5 units of the true A10 (amplitude at 10 min) with 95% Confidence Interval and allowing multiple comparisons. The sample size was calculated based on the pilot study results; the Standard Deviation in A10 was estimated at 21.9. The formula used to estimate the sample size was:

Z21-α/2 = 1.96

where σ = Standard deviation

n = Z21-α/2 σ2

d2

A power analysis of the study using R time as the variable showed a power of 99. Demographic profile, baseline hemogram, and coagulation parameters at different time intervals and in relation to different fluids administered were analyzed using analysis of variance (ANOVA) and post hoc comparison with the Bonferroni correction applied to adjust the level of significance. Log transformation was applied for the skewed data and P value was adjusted for ISS, platelets, and intraoperative blood loss. Categorical data of inhomogeneous distribution (sex, diagnosis, mode of injury, type of anesthesia, hemoglobin, TEG parameters during different time intervals with normal and abnormal values) were analyzed using Pearson Chi-square and expressed as frequency (%). Multiple comparisons (within a group and between groups) of each coagulation parameter were performed using ANOVA and data were expressed as Mean±SD. Statistical analysis of postoperative sample adjustment for duration of surgery, intraoperative blood loss, and fluids transfused was done using analysis of covariance (ANCOVA) and data expressed as Mean±SD. Pairwise comparisons with 95% confidence interval for difference and Bonferroni adjustment for multiple comparisons was used. P values of less than 0.05 were considered statistically significant. All statistical tests were performed using commercially available statistical software (SPSS for windows version 15.0 Chicago, IL, USA) and graphs were produced using Microsoft Excel for MAC 2011(version 14.1.2).

RESULTS

All 100 trauma patients completed the study according to protocol and were included in the analysis. Distribution of subjects according to the demographic profile, baseline hemogram, and coagulation parameters is summarized in Table 1. Statistically significant difference was observed in sex distribution, diagnosis of patients, and type of anesthesia administered. Statistical analysis of postoperative sample adjustment for duration of surgery, intraoperative blood loss, and fluid transfused showed non-significant pairwise comparisons of all the TEG values [Table 2].

Table 1

Distribution of subjects according to demographic profile, baseline hemogram and coagulation parameters

Table 2

Statistical analysis of post operative sample adjusting for duration of surgery, intraoperative blood loss and fluid transfused

Baseline preoperative routine coagulation and TEG parameters amongst surgical trauma patients were similar among the four groups [Figure 1]. Crystalloid (NS/RL) preloading did not influence the coagulation parameters (conventional and TEG). In all the patients preloaded with gelatin, there was a highly significant increase in PTI (14.88±0.90 vs. 13.78±3.01, P<0.001) and a significant increase in INR (1.12±0.09 vs. 1.09±0.19, P<0.05) compared to the baseline value. An increase was observed in the parameters, namely, PTI (14.17±1.14, P<0.001) and INR (1.09±0.11, P<0.05) in the postoperative period also. In the HES group, there was statistically significant increase in prothrombin time (PT) (15.70±1.51 vs. 13.74±0.75, P=0.01) and INR (1.20±0.15 vs. 1.03±0.17, P<0.001) as compared to the baseline. Rest of the variables, i.e., APTT, R time, k time, α time, A10, A20, and MA, were not influenced in any group after preloading or at the end of the surgery.

Figure 1

Various coagulation parameters among trauma surgical patients according to the period of sample withdrawn and according to the type of fluid given

In the intergroup comparisons [Figure 1], the patients preloaded with HES had a significant increase in INR (1.20±0.15 vs. 1.12±0.09, P=0.04) and R time (6.84±2.55 min vs. 4.79±1.77 min, P=0.02) as compared to the gelatin group. At the end of the surgery, the same group (HES) had significantly more R time (6.38±3.56 min, P=0.02) as compared to the gelatin group (3.9±2.1 min). After preloading with RL, the increase in PTI (14.38±1.66 vs. 15.70±1.51, P=0.02) was statistically less as compared to the HES group. Between these two groups, the fall in k time (2.16±0.98 vs. 3.94±2.6, P=0.02), rise in MA (61.94±14.08 vs. 50.11±14.10, P=0.04), and rise in A20 (56.17±14.66 vs. 43.11±14.24, P=0.05) were more in patients preloaded with RL as compared to the HES group. At the end of the surgery, the patients who had received RL had significantly higher value of MA (61.83±16.58 vs. 46.85±18.61, P=0.02) and more increase in A20 (57.43±17.00 vs. 43.54±17.31, P=0.04) as compared to the gelatin group.

Table 3 depicts the distribution of TEG parameters during different time intervals (preoperative, intraoperative, and at the end of the surgery). In all the four groups, preoperatively a statistically significant number of patients were hypercoagulable (R time <9 min (rapid initial fibrin formation), P=0.04), (k time <4 min (rapid fixed level of clot strength), P=0.02), (α time >43 (rapid rate of fibrin buildup and cross-linking, i.e., clot formation rate), P=0.03), but had MA <48 mm (poor clot strength) (P=0.05) values. These results were observed maximum in the RL group, which had the maximum frequency of patients showing rapid clot formation, propagation, and cross-linking but poor clot strength preoperatively. However, during intraoperative period, the number of patients with abnormal TEG parameters was not significant. 100% patients in the gelatin group, 84.2% patients in the NS group, 94.4% patients in the RL group, and 66.7% patients in the HES group had hypocoagulable (R time >14 min) state in the postoperative period.

Table 3

Statistical evaluation and distribution of thromboelastography parameters during different time interval

DISCUSSION

Perioperative coagulation assessment is important in the clinical setting of trauma to diagnose the cause of bleeding, guide hemostatic therapies, and predict bleeding risk in surgical interventions.[4] Different fluids have different effects on hemostasis, attributable to either dilution of clotting factors or the substance-specific effects of the plasma substitute. Both in vivo and in vitro studies have demonstrated that crystalloids have lesser effect on coagulation system than colloids.[5-8] Among the colloids, gelatin shows lesser effect than HES.[1910] However, none of the studies compares all the commonly used fluids (NS, RL, HES, and gelatin) in trauma patients who are prone to develop coagulopathy. We selected trauma patients undergoing elective surgery to have a homogenous group of population. Though the four solutions studied do not have similar effects on plasma volume expansion, we decided to use the same volume of each to remove the confounding factor of different fluid volumes infused. Various contrasting results regarding use of different fluids have been obtained in both in vivo and in vitro studies, which have been discussed in Table 4.

Table 4

Characteristic studies comparing different fluid admistration effect on thromboelastography values

Hypothermia, acidosis, and dilution from standard resuscitation can worsen the presenting coagulopathy in a trauma victim. Hence, judicious use of resuscitation fluids becomes a challenge for the treating anesthesiologist.[11] As viscoelastic techniques have been used for coagulation assessment in multiple clinical settings but their experience in trauma is limited, we undertook this study to assess the influence of various fluid preparations (gelatin, HES, RL, and NS) on coagulation using standard coagulation parameters and real-time TEG in patients undergoing elective surgery post trauma.

There is a marked male preponderance in all communities of the world among trauma victims,[12] which was statistically significant in our study as well [Table 1]. Weights of patients were statistically similar, and hence the estimated blood volumes of patients were similar across the groups [Table 1]. Viscoelastic hemostatic assays (VHA) in majority of patients with minor trauma, moderate trauma (ISS: 10-20), severe injury (ISS: 20-35), and massive tissue injury (ISS >30) showed normal, hypercoagulability, hypocoagulability, and hyperfibrinolysis, respectively.[13] In our study, preoperative ISS (Mean±SD) fell between 10 and 20, i.e., moderate injury in all the four groups, and the frequency of patients that were hypercoagulable was statistically significant, which was consistent with the report of Johannson et al.[13] Effect of individual anesthesia technique on hemostasis and coagulation parameters is controversial. The study by Huang et al. failed to demonstrate any enhancement of hemostasis or fibrinolysis postoperatively in patients undergoing arthroscopy under either general anesthesia or spinal anesthesia.[14] However, stress induced by tracheal intubation may result in catecholamine surge resulting in enhanced platelet aggregation and hence accelerated blood coagulation in other studies.[15]

Preloading with gelatin and HES caused increase in PTI and INR compared to the baseline [Figure 1], and this influence could be observed in postoperative period also in the gelatin group but not in the HES group. However, no significant effect of preloading on TEG parameters indicates that interaction of hemostatic components in vivo is not affected by the time of sampling. This observation can be explained by the fact that the dose we used for in vivo hemodilution (<40% and <28 ml/kg) was low as compared to some other studies.[16]

On intergroup comparison [Figure 1], preloading with HES delayed clot formation (increased R time and k time) and resulted in weaker clot (decreased MA) but slower clot lysis at 20 min after attaining MA, as compared with crystalloids. These results are consistent with the previous studies which state that HES 6% causes weaker clot with less stable fibrin network and less firm aggregation of platelets.[13617-19] Preloading with RL when compared to HES, on the other hand, resulted in statistically significant decrease in k time and increase in MA and A20, indicating rapid clot propagation and strong clot, but rapid fibrinolysis. Ansari et al.[20] observed that in vitro dilution of blood to 60% using lactated Ringer's/HES (130/0.4) combination produced significantly more derangement in all parameters compared with RL alone (increased clotting time and clot formation time, and decrease in maximum clot formation). A statistically significant rise in PTI after HES preloading compared to RL does not represent clot dynamics or quality. Postoperatively; gelatin caused more rapid clot formation (R time minimum) compared to HES, with weaker clot strength (lesser MA) and slower fibrinolysis (lesser A20) compared to RL. Konrad et al.[21 compared in vitro coagulation in gelatin, 6% HES (450/0.7), and RL in 33% and 66% dilutions, measuring routines laboratory and SONOCLOT (viscoelastic) variables. Hemodilution with RL tended to increase in vitro coagulability. Amongst the tested colloids, gelatin had the least impact on markers of coagulation. HES had the largest impact on markers of coagulation compared with gelatin and RL, which is similar to the results obtained by Neimi et al.[22] and Mittermayr et al.[1]

Our study demonstrates that crystalloids effect coagulation much lesser than colloids as they mainly exhibit diluting effect on coagulation system. RL is better than NS as it promotes rapid clot propagation, strong clot, and rapid fibrinolysis. Previous studies demonstrate resuscitation with RL reduces tissue hypoxia indices but does not effect the changes in fibrinogen metabolism resulting from hemorrhage.[33] In our study, HES delayed clot formation and resulted in weaker clot but slow clot lysis. Such hypocoagulability may be detrimental in patients having increased bleeding risk, e.g., severe trauma (ISS >20), who are already hypocoagulable.[13] Such an effect might be because HES is a highly branched and hydroxyethylated glucose polymer that can reduce von Willebrand (vWB) factor and interferes with fibrinogen function and polymerization.[3434] We chose HES 130/0.4, which is known to effect coagulation to the least because of its low molecular weight and low degree of substitution, amongst other starches.[35] Gelatin caused more rapid clot formation, but weakest clot strength and slowest fibrinolysis that might be due to the dilutional effect, and gelatin solutions might influence the weight and reticular network of fibrin strands and platelet function with decreased vWB factor.[3436]

One limitation of the study was that numbers in each group were relatively small, which led to inequities in sex distribution between groups and in the kinds of operations and anesthesia. Larger groups would be necessary to correct this problem. Another pitfall of our study was that our study population included patients of ISS 10-20 (moderately injured patients) undergoing elective surgeries. However, in patients with severe or life-threatening injury, high volume of fluids will be infused that would result in more pronounced hemoalterations. Besides, our study was an in vivo study; the effect of extreme hemodilution with larger volumes of fluid was not investigated upon, which might occur in patients of trauma during resuscitation. Type of surgery and type of anesthesia were the uncontrolled variables in our study. TEG may not detect platelet adhesion abnormalities such as vWB factor deficiency or drug-induced platelet inhibition.[37]

In conclusion, the choice of fluid for elective surgery does not matter, as the coagulation abnormalities observed are clinically irrelevant. However, the results might me more readily explained by observing that gelatin and HES expand the circulation more than NS or LR. Our study demonstrates that crystalloids are optimal volume expanders in trauma, with RL having beneficial effects on coagulation system (decrease in k time, increase in MA and A20), however, which is clinically irrelevant. Among the colloids, HES 6% (130/0.4) affects coagulation parameters (increase in PTI, INR, R time, k time) more than gelatin.