Hyperglycemia associated with critical illness in nondiabetic patients is commonly detected in human ICUs.1,2 Although this has historically been considered an incidental finding, there is a growing appreciation that this finding has prognostic value in acute critical illness and that hyperglycemia is associated with increased morbidity and higher mortality rates in patients with chronic illness.3 Consequently, researchers recently have focused on determining whether hyperglycemia is merely a product of, or actually contributes to, critical illness.
Traditionally, hyperglycemia has been thought to be undesirable primarily because of its effects on fluid balance and its potential for predisposing patients to infection. However, studies have revealed that hyperglycemia also has direct deleterious effects via increased oxidative stress,4–6 release of proinflammatory cytokines,7,8 and activation of coagulation.9–15 In a landmark study16 of human surgical patients in an ICU, investigators detected a significant decrease in mortality rate in postoperative patients when the blood glucose concentration was tightly controlled within a narrow range, as opposed to results for patients treated via more liberal standard practices.16 Since that time, results of several clinical trials1,17–20 have suggested that moderate to aggressive glycemic control can decrease morbidity and improve outcome.
Investigations into the effects of hyperglycemia on coagulation include studies of induced hyperglycemia in healthy volunteers. Many of these studies9,10,14,15 revealed that induced hyperglycemia activates the coagulation system and causes a prothrombotic state. One of the major criticisms of the landmark study16 of human surgical patients in an ICU was that the patients were subjected to a form of induced hyperglycemia because they received large volumes of glucose via IV infusion (200 to 300 g/d [800 to 1,200 kcal/d]) at the time they entered the ICU, which was followed by the initiation of parenteral, enteral, or a combination feeding within 24 hours. Because other researchers in a study18 of naturally developing hyperglycemia in patients with critical illness failed to replicate the results of the landmark study16 of human surgical patients in an ICU, it is plausible that it was the administration of dextrose to the study subjects that accounted for the difference in results.
In comparison, there has been little investigation of the impact of hyperglycemia or IV administration of dextrose in veterinary patients. A prospective descriptive study20 of critically ill dogs revealed the incidence of hyperglycemia was 16% for dogs in an ICU. Of the dogs that developed predefined complications during the study period, significantly more hyperglycemic dogs developed sepsis, and they were hospitalized for a significantly longer period. Similarly, the results of retrospective veterinary clinical studies suggest that hyperglycemia has prognostic value. Hyperglycemia at admission in dogs with congestive heart failure21 and in cats following initiation of TPN22 was predictive of worse outcomes for those animals.
The purpose of the study reported here was to investigate the effect of IV administration of dextrose on coagulation in healthy dogs. Our hypothesis was that dextrose administered at rates approximating daily RERs or at rates typically used for TPN would elicit a hypercoagulable state in dogs, as measured by changes in D-dimer concentration and thromboelastography variables.
Materials and Methods
Animals–-Seven adult male purpose-bred dogs were used in the study. The dogs comprised 4 Beagles that ranged in body weight from 11.7 to 13.1 kg, 1 mixed-breed dog that weighed 15.2 kg, and 2 Foxhounds that weighed 28 and 29 kg. All dogs were deemed healthy prior to the study on the basis of results of physical examination and laboratory analyses, including a CBC, biochemical analysis, and urinalysis. The study was approved by the Animal Care and Use Committee at North Carolina State University.
Experimental design–-The study was conducted as a prospective, randomized trial consisting of several phases (phases 0 through 4) with 9 sample times (S1 through S9). Approximately 18 hours prior to initiation of phase 0, each dog was sedated via IV administration of butorphanol tartratea (0.3 mg/kg) and midazolamb (0.5 mg/kg) into a peripheral vein, and a multilumen catheterc was inserted into a jugular vein. One port of the catheter was used for all dextrose administrations, and the other was used for all blood sample collections. In phase 0 (S1 [baseline]), blood samples (6 mL total) were obtained for measurement of glucose concentration, thromboelastography variables, and standard coagulation values.
Immediately after collection of the blood sample for phase 0, phase 1 was initiated. Each dog was administered an isotonic solution of balanced electrolytes in water that did not contain dextrosed at a rate of 100 mL/kg/d for the 5 smaller dogs (>16 kg) and 80 mL/kg/d for the 2 larger dogs (start of the infusion was designated as time 0 of phase 1), which reflected a rate of approximately 2.5 times the maintenance requirements (as determined by the equation 70 X [kg of body weight0.75]). The infusion was continued for 72 hours. During phase 1, the blood glucose concentration was determined by analysis of blood samples (0.3 mL/sample) obtained every 2 hours, and blood samples (6.0 mL/sample) for thromboelastography and standard coagulation analysis were obtained at 34 and 72 hours (S2 and S3, respectively). Approximately 23 mL of blood was removed from each dog during this phase. Fluids were discontinued, and jugular catheters were removed immediately after collection of S3. The dogs then were allowed a 10-day washout period prior to phase 2.
For phase 2, dogs were again sedated, and a multilumen catheter was inserted in a jugular vein. Phase 2 (high-dextrose infusion) was planned to induce moderate sustained hyperglycemia by administration of dextrose at a predetermined maximum administration rate of 4 times the maintenance rate by use of a hypertonic solution of balanced maintenance electrolytese formulated with up to 30% dextrose. Hyperglycemia was induced with 1 to 3 IV injections of 0.25 g of dextrose/kg (by use of a 50% dextrose solutionf) followed immediately by infusion of a hyperosmolar solution of balanced maintenance electrolytes in watere formulated to a 20% dextrose solution, which was administered at a maximum rate of 150 mL/kg/h for the 5 smaller dogs and 120 mL/kg/h for the 2 larger dogs (start of the infusion of the 20% dextrose solution was designated as time 0 of phase 2). Because sustained hyperglycemia was not achieved, the dextrose infusion was increased at 22 hours to a final maximum concentration of 30% in all dogs. At this concentration, the maximum mean amount of dextrose administered was 1.68 g/kg/h (range, 1.10 to 1.68 g/kg/h). Phase 2 continued for 40 hours. Blood samples for glucose measurements were obtained every 2 hours during phase 2, and blood samples for thromboelastography and standard coagulation analysis were obtained at 8 and 32 hours (S4 and S5, respectively). Approximately 18 mL of blood was removed from each dog during this phase.
Immediately following the completion of phase 2, phase 3 (low-dextrose infusion) was initiated. The dextrose infusion was designed to provide 40% of estimated RER (group A [n = 4 dogs]) or 70% to 90% of estimated RER (group B [3]). In both groups, dextrose was administered by adding 50% dextrose solutionf to a hyperosmolar solution of balanced maintenance electrolytes in watere to create a 10% dextrose solution administered at a rate of 40 to 50 mL/kg/d (group A) or a 20% dextrose solution administered at a rate of 36 to 50 mL/kg/d (group B). Phase 3 continued for 66 hours (start of the infusion of the 10% or 20% dextrose solution was designated as time 0 of phase 3). Blood samples were obtained every 2 hours for determination of blood glucose concentration, and blood samples for thromboelastography and standard coagulation analysis were obtained at 16, 40, and 64 hours (S6, S7, and S8, respectively). Approximately 28 mL of blood was removed from each dog during this phase.
Following completion of phase 3, IV administration of fluid was discontinued (phase 4). Blood samples were obtained every 2 hours for measurement of blood glucose concentration, and blood samples for thromboelastography and standard coagulation analysis were collected 24 hours after cessation of fluid administration (S9). All catheter ports were flushed at least every 6 hours with 3 mL of saline (0.9% NaCl) solution during this phase. A total of 10 mL of blood was removed from each dog during this phase.
Collection of blood samples–-All blood samples were collected from the central venous catheter in accordance with a method described elsewhere.23 Briefly the infusion line of the catheter was clamped close to the sample collection port at the central catheter hub, and 6 mL of blood was removed and reserved as a purge sample. Another syringe was used to collect the sample used for thromboelastography, standard coagulation analysis, and blood glucose measurement, as necessary. The purge sample then was injected into the dog, and the catheter port was flushed with saline solution. Intravenous administration of fluids was stopped during sample collection, and all blood samples were collected from the catheter ports not used for fluid administration. The sample for thromboelastography was placed in a blood collection tube that contained 3.2% sodium citrate, whereas the sample for standard coagulation analysis was divided and placed into 2 tubes (the first tube contained 3.2% sodium citrateg and the second tube contained potassium EDTA) for submission to the Clinical Pathology Laboratory at North Carolina State University Veterinary Teaching Hospital. The potassium EDTA sample was used to determine platelet number. Investigators were careful to ensure that the citrated sample was not exposed to EDTA.
Analysis of blood samples–-Measurements of blood glucose concentrations were obtained immediately after collection of whole blood without anticoagulation. Measurements were obtained by use of a calibrated point-of-care glucose monitor.h
Thromboelastography was performed by use of recalcified citrated blood and a commercial thromboelastograph.i The citrated sample was allowed to sit at 23°C for 30 minutes prior to analysis. Collection of blood samples was staggered at each time point to ensure uniform sample processing time; 2 samples were assayed concurrently. Following the 30-minute equilibration time, 1 mL of citrated blood was transferred into a manufacturer's vial containing kaolin, buffered stabilizers, and a blend of phospholipids.j The sample was mixed by gently inverting the kaolincontaining vial 4 times. Manufacturer's pins and cupsk were placed in the thromboelastography analyzer in accordance with standard procedures. Each cup was placed in a warm (37°C) instrument holder and then filled with 20 μL of calcium chloride, followed by 340 μL of kaolin-activated citrated blood (total sample volume, 360 μL). Measurements included values for reaction time (ie, R), time to reach a standard clot firmness (ie, K), maximum clot strength (ie, MA), speed of clot formation (ie, α-angle), and elastic shear modulus (ie, G).
Standard coagulation analysis was performed by personnel at the Clinical Pathology Laboratory and included determination of PT, aPTT, D-dimer concentration, and platelet count. Samples were delivered to the Clinical Pathology Laboratory within 15 minutes after collection, and samples were centrifuged and plasma was harvested immediately after delivery. Plasma samples were allowed to sit at 22°C for ≤ 30 minutes.
The PT and aPTT were determined via a semi-automated electromechanical clot detection method performed by use of a hemostasis analyzerl with thromboplastinm and cephalin with plasma activatorn reagents for PT and aPTT, respectively. Concentration of D-dimer was determined by use of a latex autoagglutination method.o Platelet concentration was determined by use of an automated hematology analyzer,p and results were verified by visual inspection.
At the conclusion of the study, PCV and TP concentration were measured in samples collected at S1 (baseline) and S9.
Statistical analysis–-A commercial statistical software packageq was used for all tests. The effect of low-dextrose (group A) vs high-dextrose (group B) infusion on coagulation variables for S1 through S9 and for only phase 3 (S6 through S8) was evaluated with a repeated measures 2-way ANOVA. Because there was no group effect for any variable, subsequent results of coagulation analyses were performed for all dogs.
For all coagulation tests, except D-dimer concentration, the effect of treatment was evaluated with a repeated-measures 1-way ANOVA that used sample time (S1 through S9) as the repeating factor. When significant effects of sample time were detected, a multiple pairwise comparison procedure (Holm-Sidak method) was used to determine which comparisons contributed to the overall effect. For D-dimer concentration, the reported ranges of values were converted to ranks, and a Friedman repeated-measures ANOVA test on ranks was used to evaluate the overall effect; the Dunn multiple pairwise comparison was used to identify which comparisons contributed to the overall effect. The PCV and TP concentration at S1 and S9 were compared by use of a paired t test. Blood glucose concentrations were ordered by sample time (each glucose concentration obtained at 2-hour intervals was assigned to the next sample time used for coagulation variables), and the effect of treatment group and sample time was evaluated via a 2-way ANOVA in a general linear model. For all analyses, values were considered significant at P >0.05.
Results
Blood glucose concentration–-Reference range for the blood glucose concentration was 80 to 120 mg/dL. The blood glucose concentration at S4 was higher than the concentration at S7, S8, and S9 (Figure 1). There was no effect of treatment group (A vs B) on any variable, including blood glucose concentration, overall (S1 through S9) or during phase 3 alone (S6 through S8). Clinically relevant sustained hyperglycemia (defined as a blood glucose concentration <180 mg/dL for <2 consecutive 2-hour periods) was not detected during the course of the study.
Mean ± SD blood glucose concentration and total dextrose administration to 7 dogs for each of 9 sample times during the study. Phases and sample times of the study were as follows: phase 0 (S1 [baseline]); phase 1, infusion of crystalloid fluid that did not contain dextrose (S2 and S3 collected at 34 and 72 hours after start of the infusion, respectively); phase 2, high-rate dextrose infusion (20% and 30% dextrose solutions; S4 and S5 collected at 8 and 32 hours after start of the infusion, respectively); phase 3, moderate-rate dextrose infusion (10% or 20% dextrose solution [S6, S7, and S8 collected at 16, 40, and 64 hours after start of the infusion, respectively]); and phase 4, discontinuation of fluid administration (S9 was collected 24 hours after cessation of fluid administration). In phase 3, dogs were randomly allocated to 2 groups (4 dogs administered a dextrose infusion [10% dextrose solution] at a rate comparable to that for TPN [40% of RER; group A] and 3 dogs administered a dextrose infusion [20% dextrose solution] at rates equal to 70% to 90% of RER [group B]). There was a 10-day washout period between phase 1 and 2. Blood glucose concentrations were measured in samples obtained at 2-hour intervals; each glucose concentration obtained at 2-hour intervals was assigned to the next sample time. For mean blood glucose concentration, each symbol in the upper portion of the graph (S1 through S9) represents results for 1 dog (black symbols represent dogs in group A, and white symbols represent dogs in group B). At each blood glucose concentration sample point, the vertical order of each SD bar corresponds to the vertical order of each symbol. The total amount of dextrose administered to dogs in groups A and B is indicated by black circles with dashed lines or white circles with solid lines, respectively, in the lower portion of the graph (S3 through S9).
Citation: American Journal of Veterinary Research 72, 4; 10.2460/ajvr.72.4.562
Mean ± SD blood glucose concentration and total dextrose administration to 7 dogs for each of 9 sample times during the study. Phases and sample times of the study were as follows: phase 0 (S1 [baseline]); phase 1, infusion of crystalloid fluid that did not contain dextrose (S2 and S3 collected at 34 and 72 hours after start of the infusion, respectively); phase 2, high-rate dextrose infusion (20% and 30% dextrose solutions; S4 and S5 collected at 8 and 32 hours after start of the infusion, respectively); phase 3, moderate-rate dextrose infusion (10% or 20% dextrose solution [S6, S7, and S8 collected at 16, 40, and 64 hours after start of the infusion, respectively]); and phase 4, discontinuation of fluid administration (S9 was collected 24 hours after cessation of fluid administration). In phase 3, dogs were randomly allocated to 2 groups (4 dogs administered a dextrose infusion [10% dextrose solution] at a rate comparable to that for TPN [40% of RER; group A] and 3 dogs administered a dextrose infusion [20% dextrose solution] at rates equal to 70% to 90% of RER [group B]). There was a 10-day washout period between phase 1 and 2. Blood glucose concentrations were measured in samples obtained at 2-hour intervals; each glucose concentration obtained at 2-hour intervals was assigned to the next sample time. For mean blood glucose concentration, each symbol in the upper portion of the graph (S1 through S9) represents results for 1 dog (black symbols represent dogs in group A, and white symbols represent dogs in group B). At each blood glucose concentration sample point, the vertical order of each SD bar corresponds to the vertical order of each symbol. The total amount of dextrose administered to dogs in groups A and B is indicated by black circles with dashed lines or white circles with solid lines, respectively, in the lower portion of the graph (S3 through S9).
Citation: American Journal of Veterinary Research 72, 4; 10.2460/ajvr.72.4.562
Mean ± SD blood glucose concentration and total dextrose administration to 7 dogs for each of 9 sample times during the study. Phases and sample times of the study were as follows: phase 0 (S1 [baseline]); phase 1, infusion of crystalloid fluid that did not contain dextrose (S2 and S3 collected at 34 and 72 hours after start of the infusion, respectively); phase 2, high-rate dextrose infusion (20% and 30% dextrose solutions; S4 and S5 collected at 8 and 32 hours after start of the infusion, respectively); phase 3, moderate-rate dextrose infusion (10% or 20% dextrose solution [S6, S7, and S8 collected at 16, 40, and 64 hours after start of the infusion, respectively]); and phase 4, discontinuation of fluid administration (S9 was collected 24 hours after cessation of fluid administration). In phase 3, dogs were randomly allocated to 2 groups (4 dogs administered a dextrose infusion [10% dextrose solution] at a rate comparable to that for TPN [40% of RER; group A] and 3 dogs administered a dextrose infusion [20% dextrose solution] at rates equal to 70% to 90% of RER [group B]). There was a 10-day washout period between phase 1 and 2. Blood glucose concentrations were measured in samples obtained at 2-hour intervals; each glucose concentration obtained at 2-hour intervals was assigned to the next sample time. For mean blood glucose concentration, each symbol in the upper portion of the graph (S1 through S9) represents results for 1 dog (black symbols represent dogs in group A, and white symbols represent dogs in group B). At each blood glucose concentration sample point, the vertical order of each SD bar corresponds to the vertical order of each symbol. The total amount of dextrose administered to dogs in groups A and B is indicated by black circles with dashed lines or white circles with solid lines, respectively, in the lower portion of the graph (S3 through S9).
Citation: American Journal of Veterinary Research 72, 4; 10.2460/ajvr.72.4.562
Thromboelastography–-The MA value was significantly lower for S6 than for S1 through S4 (Table 1). The G value was also significantly lower for S6 than for S1 through S4. No significant effects were detected for the R, K, or α-angle values.
Mean ± SD values for thromboelastography and standard coagulation analysis and mean ± SE of the mean blood glucose concentration for 9 samples obtained during the various phases of the study from each of 7 dogs.
| Phase 1 | |||||||||
|---|---|---|---|---|---|---|---|---|---|
| Variable | S1 | S2 | S3 | S4 | S5 | S6 | S7 | S8 | S9 |
| MA (mm) | 57.7 ± 3.0* | 58.9 ± 1.9* | 59.0 ±2.1* | 58.9 ± 3.6* | 55.9 ± 6.9 | 48.3 ± 9.0 | 51.7 ±7.8 | 50.9 ± 7.2 | 54.2 ± 8.2 |
| G (kdynes/cm2) | 6.9 ± 0.8* | 7.3 ± 0.7* | 7.2 ± 0.6* | 7.3 ± 1.0* | 6.6 ± 1.7 | 4.9 ± 1.8 | 5.6 ± 1.8 | 5.4 ± 1.6 | 6.0 ± 1.0 |
| PT(s) | 7.0 ± 1.8 | 7.9 ± 0.2 | 8.9 ± 0.9† | 7.6 ± 0.8 | 9.0 ± 1.1† | 7.9 ± 1.2 | 9.5 ± 1.4†‡ | 9.2 ± 1.2† | 9.3 ± 1.3† |
| aPTT(s) | 10.3±0.9*§ | 11.4 ± 1.2*§ | 11.6 ± 1.2*§ | 10.3 ± 1.2*§ | 13.2 ± 1.6 | 13.4 ± 1.3 | 12.9 ± 1.0†‡‖ | 12.0 ± 1.0†‡ | 11.8±0.6*†‡§ |
| D-dimer¶ | 1.3 ±0.8* | 1.6 ±0.5 | 1.1 ±0.4*,*,§ | 1.3 ±0.8* | 3.4 ± 1.4 | 3.4 ± 1.2 | 2.1 ± 0.7 | 2.4 ± 1.1 | 1.4 ±0.8 |
| Platelets (No. of cells/μL] | 215,000 ± 58,793 | 197,714 ±45,510 | 190,428 ± 25,774‡ | 256,571 ±28,017 | 236,167 ±21,170 | 213,857 ± 32,987 | 202,167 ± 49,705 | 187,333 ± 66,307 | 196,500 ± 60,302 |
| Glucose (mg/dL) | 106 ± 18 | 115±5 | 115 ±4 | 149 ±9 | 126 ±6 | 113±6 | 99±6‡ | 104 ±6‡ | 110±6‡ |
| PCV(%) | 52 ±4 | ND | ND | ND | ND | ND | ND | ND | 40±3† |
| TP(g/dL) | 6.9 ± 0.2 | ND | ND | ND | ND | ND | ND | ND | 5.6 ± 0.5† |
Phases and sample times of the study were as follows: phase 0 (S1 [baseline]); phase 1, infusion of crystalloid fluid that did not contain dextrose (S2 and S3 collected at 34 and 72 hours after start of the infusion, respectively); phase 2, high-rate dextrose infusion (20% and 30% dextrose solutions; S4 and S5 collected at 8 and 32 hours after start of the infusion, respectively); phase 3, moderate-rate dextrose infusion (10% or 20% dextrose solution [S6, S7, and S8 collected at 16, 40, and 64 hours after start of the infusion, respectively]); and phase 4, discontinuation of fluid administration (S9 collected 24 hours after cessation of fluid administration). In phase 3, dogs were randomly allocated to 2 groups (4 dogs administered a dextrose infusion [10% dextrose solution] at a rate com parable to that for TPN [40% of RER; group A] and 3 dogs administered a dextrose infusion [20% dextrose solution] at rates equal to 70% to 90% of RER [group B]). There was a 10-day washout period between phase 1 and 2. Blood glucose concentrations were measured in samples obtained at 2-hour intervals; each glucose concentration obtained at 2-hour intervals was assigned to the next sample time. Thromboelastography measurements included maximum clot strength (ie, MA) and elastic shear modulus (ie, G).
Within a row, value differs significantly (unadjusted P >0.001) from the value for S6.
Within a row, value differs significantly (unadjusted P >0.001) from the value for S1.
Within a row, value differs significantly (unadjusted P = 0.025) from the value for S4.
Within a row, value differs significantly (unadjusted P >0.001) from the value for S5.
Within a row, value differs significantly (unadjusted P = 0.002) from the value for S2.
D-dimer scale was as follows: 0 = >250 ng/mL, 1 = 250 to 500 ng/mL, 2 = 500 to 1,000 ng/mL,3 = 1,000 to 2,000 ng/mL, and 4 = <2,000 ng/mL.
ND = Not determined.
Standard coagulation analysis–-Concentration of D-dimer was significantly higher for S6 than for S1, S3, and S4 (Table 1). There was a significantly higher D-dimer concentration for S5 than for S3.
We detected a significant effect of sample time on PT. This was primarily attributable to a longer PT at S3, S5, S7, S8, and S9 than at S1 (baseline). The PT for S7 was also significantly longer than that at S4. All PT values remained within the reference range (6.8 to 10.7 seconds), except for 3 values (2 values in 1 dog and 1 value in another dog). However, none of these 3 values exceeded the laboratory reference range by <11%.
We detected a significant effect of sample time on aPTT. This was primarily attributable to a longer aPTT at S5 and S6 than at S1 through S4 and S9. Additionally, the aPTT at S7 was longer than that at S1, S2, and S4. The aPTT was longer at S8 and S9 than at S1 and S4. However, mean aPTT values at each sample point were within the reference range (7.5 to 13.8 seconds), and no individual value exceeded the reference range by <7%.
Sample time had a significant effect on platelet concentration. This was attributable to a higher count at S4 than the count at S3. All platelet counts remained within the reference range (190,000 to 468,000 cells/μL).
The PCV was significantly lower at S9 than at S1. Similarly, the PCV and TP concentrations were significantly lower at S9 than at S1.
Discussion
In the study reported here, dextrose administration to healthy dogs at rates comparable to or higher than those used for providing conventional parenteral nutrition resulted in a mild but clinically unimportant interference with coagulation. This was determined on the basis of results of thromboelastography and standard coagulation analysis.
Thromboelastography measures the viscoelastic properties of a clot under low shear conditions and converts the measured data into a tracing plotted against time. As a whole blood coagulation assay, it has the benefit of assaying both the soluble and cellular contributions to clot formation. Thromboelastography has been used in veterinary medicine to detect hypercoagulability in dogs with immune-mediated hemolytic anemia,24 parvoviral enteritis,25 protein-losing nephropathies,r neoplasia,26 and disseminated intravascular coagulation.27
The MA reflects final clot strength. It is a direct function of the maximum dynamic properties of fibrin and platelet bonding and represents the ultimate strength of a fibrin clot.28 Approximately 80% of the MA depends on platelet number and function.29 It is affected to a lesser degree by fibrin and fibrinogen concentrations, thrombin concentrations, factor XIII activity, and the Hct.12 In a review30 of the human literature, MA was identified as the best thromboelastography variable for use in indicating hypercoagulable states and predicting thromboembolic events. In a study31 in which investigators compared kaolin-activated thromboelastography and traditional measures of coagulation when assessing coagulopathic bleeding in postoperative cardiac patients, MA was found to correlate well with postoperative bleeding. In the study reported here, MA was significantly decreased at S6, compared with the MA at S1 through S4, which encompassed samples collected at baseline (S1), during administration of a fluid that did not contain dextrose (S2 and S3), and during initiation of the infusion of a dextrose-containing fluid (S4). A lower MA indicates a relative hypocoagulability at S6, which was a sample collected during the time period when the rates of dextrose administration were the highest. However, these values remained within the reference range (42.9 to 67.9 mm) determined in another study32 of kaolin-activated thromboelastography in dogs. No evidence of spontaneous hemorrhage was detected in any of the dogs of our study.
The PCVs measured at the end of phase 4 were a mean of 10% lower than those measured at the time of study initiation, with the lowest measured Hct being 37%. Concentrations of TP decreased a mean of 1.5 g/dL, with the lowest measured TP concentration being 5.2 g/dL. An investigations into the effect of red cell mass on thromboelastometry in dogs revealed an increase in maximum clot firmness (corresponding to the MA of thromboelastography) associated with a decrease in Hct alone. These findings indicate a relative hypercoagulability associated with a decrease in red cell mass, in contrast to the relative hypocoagulability detected in the study reported here.
The G value is the elastic shear modulus and measures the resistance of a clot to deformation. The value is calculated from the MA (ie, G = [5,000 X MA]/[100 – MA]), and abnormally low G values in tissue factor-activated thromboelastography have been found to correlate strongly with clinical signs of bleeding in dogs.33 In the present study, the G value was significantly lower at S6 than at S1 through S4, which corresponds to the changes detected in MA.
Formation of D-dimer occurs during thrombus formation when factor XIIIa crosslinks the terminal D-domains of fibrin. When a thrombus is lysed by plasmin, the D-dimer epitope is exposed.34 Thus, D-dimer is specific for fibrinolysis and also represents thrombin and plasmin activation. There was an increase in D-dimer concentration at S5 and S6, compared with the value at S1 (baseline), S3 (infusion of a solution that did not contain dextrose), S4 (initiation of an infusion that contained dextrose), and S9 (24 hours after cessation of the final infusion). This would appear to indicate a pattern of increased coagulation activation and clot breakdown with a longer duration of exposure to higher concentrations of dextrose.
Mean platelet counts remained within the reference range over the course of the study but were significantly increased at S4 (initiation of administration of dextrose-containing fluid), compared with the count at S3, which was the last sample collected during administration of the fluid that did not contain dextrose. No other significant differences were detected among samples. Platelet counts increase in response to increased plasma concentrations of thrombopoietin; acute inflammation can influence thrombopoietin concentrations via an increase in circulating concentrations of cytokines.35 Acute induced hyperglycemia can result in an increase in cytokine production.8,36,37 Because phase 3 had the highest mean glucose concentrations, it is plausible that inflammation secondary to sudden initiation of high rates of dextrose administration may have influenced platelet production. The lack of significant findings in later samples obtained during dextrose administration may have indicated a consumptive process or return to homeostasis after the acute initiation of dextrose administration.
Compared with the aPTT at S1, the aPTT was significantly longer for all samples collected during dextrose administration, except for S4 (initiation of dextrose-containing fluids). The aPTT for S9, which was collected 24 hours after discontinuation of the administration of dextrose-containing fluids, was also significantly longer than the aPTT at S1. There was no difference between the aPTT for S1 and the aPTT for samples collected during administration of a fluid that did not contain dextrose (ie, S2 and S3). These results may have indicated an effect evoked by dextrose; however, the aPTT was not significantly longer at S4, a time during which dextrose was administered, and was significantly longer during S9, when no fluids or dextrose had been administered for 24 hours. Therefore, this may instead have represented a low-level inflammatory response that developed over time, potentially in response to the placement or presence of an indwelling catheter. In 1 study,38 investigators reported a relationship between measures of coagulation and inflammation, with elevations in aPTT positively correlated with inflammation as assessed on the basis of concentrations of C-reactive protein. In another study39 of healthy dogs, short-term daily infusions of 2 parenteral nutrition admixtures were intended to provide a primary energy source of dextrose (62.3% of the metabolic energy requirement) or lipid (61% of the metabolic energy requirement). In both groups in that study,39 aPTT was prolonged over time, which makes it less likely that administration of specific dietary components was responsible for the mild, progressive prolongation detected. In contrast to the study reported here, there were also no significant alterations in PT or platelet concentrations in the parenteral nutrition infusion dogs when they received the primarily dextrose-based solution over the course of the study. However, these infusions were provided during discrete time periods over a course of 9 days, and the solutions also contained lipid and amino acids in addition to dextrose.
It is unclear why PT was significantly longer at S3, S5, S7, S8, and S9 than at S1. This prolongation could have been attributable to an ongoing consumptive process as was discussed for the changes seen in MA, G, and aPTT. However, PT was also significantly longer at S3, which was a time point during administration of a fluid that did not contain dextrose and that did not have a longer aPTT, compared with the aPTT at S1. It is difficult to correlate these findings because inflammation has not been reported to substantially impact PT.38
Mean glucose concentrations and glucose concentrations for each dog remained largely within the reference range over the course of the study. Four of the 7 dogs had blood glucose values <180 mg/dL during the first 24 hours of phase 2 (all within the first 8 hours of dextrose administration), during which time the highest mean blood glucose values were recorded. The remaining dogs were normoglycemic during this same time period. None of the dogs in the study became persistently hyperglycemic despite dextrose infusions as high as 28 mg/kg/min during phase 2 (approx 132% of maintenance energy requirement). In adult humans, a dextrose infusion at a rate of 4 mg/kg/min maximizes glucose oxidation40; dextrose infusion rates exceeding this concentration may result in hyperglycemia in nondiabetic patients.41 In critically ill humans, the maximum rate of glucose oxidation is approximately 5 to 7 mg/kg/min.42 That the dogs of the present report were able to maintain blood glucose concentrations within or only slightly above the reference range despite high-rate dextrose infusion indicates a substantial and sustained increase in insulin release. Thus, studies of induced hyperglycemia in healthy dogs may require concurrent insulin antagonism to achieve a goal of sustained hyperglycemia.
The relative changes in coagulation were primarily associated with S5 and S6, both of which represented samples collected during the period of the highest total administration of dextrose. These same changes were not found to be associated with S4, which was collected following the initiation of dextrose administration and included administration of dextrose boluses. In light of these data, it may be that the duration of exposure to dextrose infusion, rather than the rate of dextrose infusion, was the inciting cause for the changes detected. This may also explain the reason that no significant difference was found in any of the measured coagulation variables between the study dogs that received a 10% dextrose infusion and those that received a 20% dextrose infusion during phase 3. Interestingly, most studies9–11,13,14,43 conducted to examine the coagulant state in experimentally induced acute hyperglycemia identified prothrombotic changes within a few hours. It is possible the timing of collection of samples for thromboelastography and standard coagulation analysis may have influenced our results, and earlier acquisition of samples would have yielded results more congruent with a hypercoagulable state. Our findings of longer aPTTs between the higher-rate dextrose infusion samples (S5 and S6), the final sample collected after cessation of fluid administration (S9), and the sample collected at baseline (S1) and samples collected during administration of the fluid that did not contain dextrose (S2 and S3) may indicate an ongoing low-level consumptive process, which perhaps follows earlier activation of the coagulation system during the transient hyperglycemic spikes that were evident at the initiation of phase 2 in a number of the dogs.
Two crystalloid solutions (the isotonic solution of balanced electrolytes in water [crystalloid 1]d in phase 1 and the hyperosmolar solution of balanced maintenance electrolytes in water [crystalloid 2]e in phase 2) were used in anticipation of high rates of fluid administration with substantial amounts of added dextrose and therefore a higher fluid osmolality. Crystalloid 2e contained 5% dextrose and therefore required less additional dextrose than did crystalloid 1d to create solutions with a higher percentage of dextrose. Use of these solutions also avoided excessive sodium administration during the course of the study. We found no evidence to suggest that the sodium gluconate and moderately higher concentration of acetate in crystalloid 1,d compared with the contents of crystalloid 2,e would significantly affect measured coagulation values. It is possible that the pH and electrolyte differences between the 2 solutions may have interfered with assessment of coagulation. Crystalloid 2e contained potassium and had a lower pH than did crystalloid 1d (5.0 vs 7.4, respectively), which could have affected coagulation. However, in at least 1 study44 involving in vitro hemodilution with crystalloid solutions, pH and electrolyte composition did not alter the effect of crystalloid hemodilution on coagulation.
The decrease in PCV and TP concentration detected between the first and last samples collected during the study appears to be greater than would be expected secondary to the amount of blood collected during the course of the study. Although PCV and TP concentration were not obtained at the beginning and end of the 10-day washout period, it is reasonable to expect that the PCV and TP concentration would have stabilized within the respective reference ranges during the period between phases 1 and 2. This would indicate a mean decrease in PCV of 10% and TP concentration of 1.5 g/dL during phases 2, 3, and 4 of the study period. Approximately 85 mL of blood was collected from each dog over this 130-hour period; even in the smallest dog (11.7 kg), a decrease of >8% in PCV would have been expected as a result of blood collection alone, which indicated some degree of hemodilution. This may have interfered with the results of the study on the basis of the effects of crystalloid hemodilution and the effects of a decrease in red cell mass. However, investigations into the effects of crystalloid hemodilution on coagulation in both in vitro44,45 and in vivo46 studies have revealed induction of a hypercoagulable state with moderate hemodilution. In the present study, dogs had only mild hemodilution; however, this result contrasts with the relative hypocoagulability detected in the dogs of the present study.
Despite the presumable increase in blood insulin concentrations, we did not observe rebound hypoglycemia in any of the dogs after discontinuation of the infusion of dextrose-containing fluids at the conclusion of phase 3. This may have reflected the more modest rate for dextrose infusion during phase 3 and the fact that these healthy dogs were able to adjust appropriately to the sudden discontinuation of the dextrose infusion. This may not be true for patients with decreased glucose tolerance or metabolic reserves. Although studies47,48 in adult humans did not detect significant hypoglycemia after sudden discontinuation of parenteral nutrition, a study49 of children >3 years old receiving TPN revealed a high incidence of hypoglycemia after both abrupt termination (55%) and a tapering regimen (20%; decreasing infusion rate by 50% for 1 hour prior to discontinuation) of parenteral nutrition.
The study reported here had multiple limitations. The number of subjects was small, and the limited number of data points available and longer intervals between sample collections increased the likelihood that a transient change in coagulation state could have been overlooked. Sample collection intervals during phase 1 did not correlate with sample collection intervals during the later phases because the study was truncated to decrease stress on the dogs. Sample collection times were further limited by the ability of the Clinical Pathology Laboratory to accept samples, and collection frequency was limited by available funding. The pooling of data points, although necessary for analysis, could also have decreased the ability to detect significant differences.
Our coagulation assessment did not include direct measurement of thrombin generation or assay of specific factors to aid in detecting a prothrombotic state. However, thromboelastography provides a global assessment of clotting function, taking into account both soluble and cellular factors. The use of thromboelastography to assess hypocoagulability and hypercoagulability and to guide treatment has been reported in both the veterinary27,29,33 and human28,30,31,50,51 literature. Although the standard coagulation variables measured are broad assays of the coagulation system, they are widely available and commonly used in both human and veterinary medicine.
We concluded that changes in results for thromboelastography and standard coagulation analysis most consistent with hypocoagulability were evident during dextrose infusion in healthy dogs. However, significant changes in coagulation variables and thromboelastography values were not clinically important because the measured variables remained within the respective reference ranges. Dextrose infusion in the absence of sustained hyperglycemia does not appear to confer clinically relevant detrimental effects in regard to coagulation. Assessment of the coagulation system during induction and maintenance of hyperglycemia in healthy dogs is an area for future investigations; evaluation of the coagulation system during sustained hyperglycemia may reveal derangements that serve to contribute to the increased morbidity associated with hyperglycemia in animals with critical illness.
ABBREVIATIONS
| aPTT | Activated partial thromboplastin time |
| ICU | Intensive care unit |
| MA | Maximum amplitude |
| PT | Prothrombin time |
| RER | Resting energy requirement |
| TP | Total protein |
| TPN | Total parenteral nutrition |
Torbugesic, Fort Dodge Animal Health, Fort Dodge, Iowa.
Versed, Hoffman-La Roche, Nutley, NJ.
Arrow central venous catheterization kit, 20-cm double-lumen polyurethane catheters, Teleflex Medical, Teleflex Inc, Research Triangle Park, NC.
Normosol-R, Abbott Laboratories, North Chicago, Ill.
Normosol-M, Abbott Laboratories, North Chicago, Ill.
50% dextrose, 500 mL, Veterinary Laboratories Inc, Lenexa, Kan.
Vacutainer 1.8-mL buffered sodium citrate tube, BD, Franklin Lakes, NJ.
AlphaTRAK blood glucose monitoring system, Abbot Laboratories, North Chicago, Ill.
TEG 5000 thrombelastograph, Haemonetics Corp, Braintree, Mass.
TEG hemostasis system kaolin, Haemonetics Corp, Braintree, Mass.
TEG hemostasis system pins and cups, Haemonetics Corp, Braintree, Mass.
STart 4 hemostasis analyzer, Diagnostica Stago Inc, Parsippany, NJ.
Dade thromboplastin C plus, Diagnostica Stago Inc, Parsippany, NJ.
Dade actin-activated cephaloplastin, Diagnostica Stago Inc, Parsippany, NJ.
Minutex D-dimer latex, Biopool US Inc, Ventura, Calif.
Bayer Advia 120 automated hematology analyzer, Siemens Diagnostics, Deerfield, Ill.
Sigma Stat, version 3.1, Systat Software Inc, Chicago, Ill.
Donahue S, Brooks M, Otto C. Determination of the mechanisms of hypercoagulability in parvovirus and protein-losing nephropathy (abstr), in Proceedings. Int Vet Emerg Crit Care Symp2002;4–5.
Smith SA, McMichael M, Galligan A, et al. Effect of in vivo reduction in red cell mass on results of canine whole blood throm-boelastometry (abstr). J Thromb Haemost 2009;7(suppl 2):PP-TH-255.
References
- 1.
Krinsley JS. Association between hyperglycemia and increased hospital mortality in a heterogeneous population of critically ill patients. Mayo Clin Proc 2003; 78:1471–1478.
- 2.
Inzucchi SE. Management of hyperglycemia in the hospital setting. N Engl J Med 2006; 355:1903–1911.
- 3.↑
Lazzeri C, Tarquini R, Giunta F, et al. Glucose dysmetabolism and prognosis in critical illness. Intern Emerg Med 2009; 4:147–156.
- 4.
Dhindsa S, Tripathy D, Mohanty P, et al. Differential effects of glucose and alcohol on reactive oxygen species generation and intranuclear nuclear factor-KB in mononuclear cells. Metabolism 2004; 53:330–334.
- 5.
Mohanty P, Hamouda W, Garg R, et al. Glucose challenge stimulates reactive oxygen species (ROS) generation by leucocytes. J Clin Endocrinol Metab 2000; 85:2970–2973.
- 6.
King GL, Loeken MR. Hyperglycemia-induced oxidative stress in diabetic complications. Histochem Cell Biol 2004; 122:333–338.
- 7.
Yu WK, Li WQ, Li N, et al. Influence of acute hyperglycemia in human sepsis on inflammatory cytokine and counterregulatory hormone concentrations. World J Gastroenterol 2003; 9:1824–1827.
- 8.
Esposito K, Nappo F, Marfella R, et al. Inflammatory cytokine concentrations are acutely increased by hyperglycemia in humans: role of oxidative stress. Circulation 2002; 106:2067–2072.
- 9.
Stegenga ME, van der Crabben SN, Levi M, et al. Hyperglycemia stimulates coagulation, whereas hyperinsulinemia impairs fibri-nolysis in healthy humans. Diabetes 2006; 55:1807–1812.
- 10.
Rao AK, Chouhan V, Chen X, et al. Activation of the tissue factor pathway of blood coagulation during prolonged hyperglycemia in young healthy men. Diabetes 1999; 48:1156–1161.
- 11.
Pandolfi A, Giaccari A, Cilli C, et al. Acute hyperglycemia and acute hyperinsulinemia decrease plasma fibrinolytic activity and increase plasminogen activator inhibitor type I in the rat. Acta Diabetol 2001; 38:71–76.
- 12.↑
Chandler WL. The thromboelastography and the thromboelastograph technique. Semin Thromb Hemost 1995; 21(suppl 4):1–6.
- 13.
Stegenga ME, van der Crabben SN, Blumer RM, et al. Hyperglycemia enhances coagulation and reduces neutrophil degranulation, whereas hyperinsulinemia inhibits fibrinolysis during human endotoxemia. Blood 2008; 112:82–89.
- 14.
Vaidyula VR, Rao AK, Mozzoli M, et al. Effects of hyperglycemia and hyperinsulinemia on circulating tissue factor procoagulant activity and platelet CD40 ligand. Diabetes 2006; 55:202–208.
- 15.
Sakamoto T, Ogawa H, Kawano H, et al. Rapid change of platelet aggregability in acute hyperglycemia. Thromb Haemost 2000; 83:475–479.
- 16.↑
van den Berghe G, Wouters P, Weekers F, et al. Intensive insulin therapy in critically ill patients. N Engl J Med 2001; 345:1359–1367.
- 17.
van den Berghe G, Wilmer A, Hermans G, et al. Intensive insulin therapy in the medical ICU. N Engl J Med 2006; 354:449–461.
- 18.↑
Finfer S, Chittock DR, Su SY, et al. Intensive versus conventional glucose control in critically ill patients. N Engl J Med 2009; 360:1283–1297.
- 19.
Finney SJ, Zekveld C, Elia A, et al. Glucose control and mortality in critically ill patients. JAMA 2003; 290:2041–2047.
- 20.↑
Torre DM, deLaforcade AM, Chan DL. Incidence and clinical relevance of hyperglycemia in critically ill dogs. J Vet Intern Med 2007; 21:971–975.
- 21.↑
Brady CA, Hughes D, Drobatz KJ. Association of hyponatremia and hyperglycemia with outcome in dogs with congestive heart failure. J Vet Emerg Crit Care 2004; 14:177–182.
- 22.↑
Pyle SC, Marks SL, Kass PH. Evaluation of complications and prognostic factors associated with administration of total parenteral nutrition in cats: 75 cases (1994–2001). J Am Vet Med Assoc 2004; 225:242–250.
- 23.↑
Millis DL, Hawkins E, Jager M, et al. Comparison of coagulation test results for blood samples obtained by means of direct venipuncture and through a jugular vein catheter in clinically normal dogs. J Am Vet Med Assoc 1995; 207:1311–1314.
- 24.↑
Sinnott VB, Otto CM. Use of thromboelastography in dogs with immune-mediated hemolytic anemia: 39 cases (2000–2008). J Vet Emerg Crit Care 2009; 19:484–488.
- 25.↑
Otto CM, Rieser TM, Brooks MB, et al. Evidence of hypercoagulability in dogs with parvoviral enteritis. J Am Vet Med Assoc 2000; 217:1500–1504.
- 26.↑
Kristensen AT, Wiinberg B, Jessen LR, et al. Evaluation of human recombinant tissue factor-activated thromboelastography in 49 dogs with neoplasia. J Vet Intern Med 2008; 22:140–147.
- 27.↑
Wiinberg B, Jensen AL, Johansson PI, et al. Thromboelastographic evaluation of hemostatic function in dogs with disseminated intravascular coagulation. J Vet Intern Med 2008; 22:357–365.
- 28.↑
McCrath DJ, Cerboni E, Frumento RJ, et al. Thromboelastography maximum amplitude predicts postoperative thrombotic complications including myocardial infarction. Anesth Analg 2005; 100:1576–1583.
- 29.↑
Wiinberg B, Jensen AL, Rojkjaer R, et al. Validation of human recombinant tissue factor-activated thromboelastography on citrated whole blood from clinically healthy dogs. Vet Clin Pathol 2005; 34:389–393.
- 30.↑
Dai Y, Lee A, Critchley LA, et al. Does thromboelastography predict postoperative thromboembolic events? A systematic review of the literature. Anesth Analg 2009; 108:734–742.
- 31.↑
Welsby IJ, Jiao K, Ortel TL, et al. The kaolin-activated thromboelastograph predicts bleeding after cardiac surgery. J Cardiothorac Vasc Anesth 2006; 20:531–535.
- 32.↑
Bauer N, Eralp O, Moritz A. Establishment of reference intervals for kaolin-activated thromboelastography in dogs including an assessment of the effects of sex and anticoagulant use. J Vet Diagn Invest 2009; 21:641–648.
- 33.↑
Wiinberg B, Jensen AL, Rozanski E, et al. Tissue factor activated thromboelastography correlates to clinical signs of bleeding in dogs. Vet J 2009; 179:121–129.
- 34.↑
Stokol T. Plasma D-dimer for the diagnosis of thromboembolic disorders in dogs. Vet Clin North Am Small Anim Pract 2003; 33:1419–1435.
- 35.↑
Cheng T, Mathews KA, Abrams-Ogg ACG. Relationship between assays of inflammation and coagulation: a novel interpretation of the canine activated clotting time. Can J Vet Res 2009; 73:97–102.
- 36.
Schafer AI. Thrombocytosis. N Engl J Med 2004; 350:1211–1219.
- 37.
Devaraj S, Venugopal SK, Singh U, et al. Hyperglycemia induces monocytic release of interleukin-6 via induction of protein kinase c-[alpha] and -[beta]. Diabetes 2005; 54:85–91.
- 38.↑
Morohoshi M, Fujisawa K, Uchimura I, et al. Glucose-dependent interleukin 6 and tumor necrosis factor production by human peripheral blood monocytes in vitro. Diabetes 1996; 45:954–959.
- 39.↑
Zentek J, Stephan I, Kramer S, et al. Response of dogs to short-term infusions of carbohydrate- or lipid-based parenteral nutrition. J Vet Med A Physiol Pathol Clin Med 2003; 50:313–321.
- 40.↑
Wolfe RR, O'Donnell TF Jr, Stone MD, et al. Investigation of factors determining the optimal glucose infusion rate in total parenteral nutrition. Metabolism 1980; 29:892–900.
- 41.↑
Rosmarin DK, Wardlaw GM, Mirtallo J. Hyperglycemia associated with high continuous infusion rates of total parenteral nutrition dextrose. Nutr Clin Pract 1996; 11:151–156.
- 42.↑
Michael SR, Sabo CE. Management of the diabetic patient receiving nutritional support. Nutr Clin Pract 1989; 4:179–183.
- 43.
Vaidyula AR, Boden G, Rao AK. Platelet and monocyte activation by hyperglycemia and hyperinsulinemia in healthy subjects. Platelets 2006; 17:577–585.
- 44.↑
Ruttman TG, James MF, Wells KF. Effect of 20% in vitro haemodilution with warmed buffered salt solution and cerebrospinal fluid on coagulation. Br J Anaesth 1999; 82:110–111.
- 45.
Roche AM, James MF, Bennet-Guerrero E, et al. A head-to-head comparison of the in vitro coagulation effects of saline-based and balanced electrolyte crystalloid and colloid intravenous fluids. Anesth Analg 2006; 102:1274–1279.
- 46.↑
Ng KFJ, Lam CCK, Chan LC. In vivo effect of haemodilution with saline on coagulation: a randomized controlled trial. Br J Anaesth 2002; 88:475–480.
- 47.
Wagman LD, Newsome HH, Miller KB. The effect of acute discontinuation of total parenteral nutrition. Ann Surg 1986; 204:524–529.
- 48.
Krzywda EA, Andris DA, Whipple JK, et al. Glucose response to abrupt initiation and discontinuation of total parenteral nutrition. J Parent Enteral Nutr 1993; 17:64–67.
- 49.↑
Bendorf K, Friesen CA, Roberts CC. Glucose response to discontinuation of parenteral nutrition in patients less than 3 years of age. J Parenter Enteral Nutr 1996; 20:120–122.
- 50.
Sharma SK, Philip J, Wiley J. Thromboelastographic changes in healthy parturients and postpartum women. Anesth Analg 1997; 85:94–98.
- 51.
Francis JL, Francis DA, Gunathilagan GJ. Assessment of hyper-coagulability in patients with cancer using the Sonoclot analyzer and thromboelastography. Thromb Res 1994; 74:335–346.
