This work was performed at New Bolton Center, University of Pennsylvania School of Veterinary Medicine. Portions of these data were presented at the ACVIM Forum in 2009.
Comparison of Hematologic and Biochemical Results on Blood Obtained by Jugular Venipuncture as Compared with Intravenous Catheter in Adult Horses
Article first published online: 24 AUG 2010
Copyright © 2010 by the American College of Veterinary Internal Medicine
Journal of Veterinary Internal Medicine
Volume 24, Issue 6, pages 1462–1466, November/December 2010
How to Cite
May, M.L., Nolen-Walston, R.D., Utter, M.E. and Boston, R.C. (2010), Comparison of Hematologic and Biochemical Results on Blood Obtained by Jugular Venipuncture as Compared with Intravenous Catheter in Adult Horses. Journal of Veterinary Internal Medicine, 24: 1462–1466. doi: 10.1111/j.1939-1676.2010.0582.x
- Issue published online: 3 NOV 2010
- Article first published online: 24 AUG 2010
- Submitted March 18, 2010; Revised June 7, 2010; Accepted July 1, 2010.
- Blood collection;
Background: During hospitalization, horses typically undergo frequent blood sampling for diagnostic testing and monitoring. The need for numerous samples in hospitalized horses makes acquisition from an intravenous catheter (IVC) both convenient and less stressful to the patient.
Hypothesis: We hypothesized that there would be no significant difference in the plasma chemistry and CBC variables from blood samples obtained from a jugular catheter as compared with direct jugular venipuncture.
Animals: Fifty adult hospitalized horses; 25 receiving constant rate crystalloid therapy, and 25 receiving low volume IV medication.
Methods: This study was conducted using a prospective, blinded, cross-over design. Samples were obtained sequentially by direct venipuncture of the jugular vein and aspiration from an IVC in the contralateral vein after an appropriate presample of blood was obtained and discarded. Samples were submitted for blinded analysis including CBC, plasma chemistry analysis, stall side plasma glucose concentration, PCV, and total protein concentration. Data obtained were analyzed using a Student's t-test with compensation for unequal variances between the 2 groups. Analyses were Bonferroni corrected for a 5% 2-tailed hypothesis test.
Results: There were no statistically significant or clinically relevant differences associated with sampling method (venipuncture versus catheter) regardless of fluid administration status in any of the 24 analytes measured.
Conclusions and Clinical Importance: Blood samples obtained by IVC have clinically equivalent values to those taken by direct venipuncture in commonly performed analyses. Additional investigation is warranted to establish if this technique is associated with increased complications such as phlebitis or bacteremia.
Patients in veterinary hospitals frequently are subjected to multiple blood sampling to allow for monitoring of serial clinical results or to obtain data relevant to changes in clinical status. Common practice at many institutions dictates the use of direct venipuncture at time of sampling despite the presence of an indwelling intravenous catheter (IVC). This practice arises from concerns that blood drawn from an IVC will either be hemolyzed and cause false electrolyte and hematology results or that the blood will be diluted by IV fluids or drugs. Repeat venipuncture is not only potentially harmful to the sampled vessel, but may increase anxiety and pain experienced by the patient, and, in fractious animals, result in potential injuries to the nursing staff. In patients with coagulopathies, drawing directly from the catheter would provide the added benefit of prevention of prolonged bleeding at the venipuncture site or thrombosis of the sampled vessel. The use of an existing IVC for repeat blood sampling may preserve vessel integrity in horses with limited venous access and minimize associated risk to the patient and personnel.
Differences in blood sampling methods and their effect on laboratory results have been analyzed repeatedly in the human medical literature. Specifically, sampling via direct venipuncture versus from a peripheral venous catheter has been found to have no effect on the accuracy of coagulation profiles1 or electrolyte results,2,3 and causes minimal hemolysis in samples drawn from large bore catheters.4 These finding were consistent even after fluid administration through the sampled IVC.5 In pediatric patients, blood sampling from an IVC was not only feasible and effective but it was not associated with any significant increase in catheter occlusion, infection, or mechanical complication rates.6 Presample blood draws have been shown to be important for accuracy of coagulation parameters and the volume withdrawn should be twice the dead space of the IVC and extension set to prevent discrepant results.7 Similarly, smaller catheter gauge was associated with increased hemolysis of samples, which could potentially impact chemistry results.4,8
Limited data are available in the veterinary literature on sampling techniques and their effect on hematology parameters. A single study in dogs found that there were no differences in coagulation times in samples drawn from a jugular IVC after a presample was discarded.9 In cats, significant differences in serum potassium, albumin, and protein concentration totals were found in blood obtained from SC vascular ports as compared with direct venipuncture, which were attributed to mild hemolysis.10 To our knowledge, no study in horses has examined CBC and plasma biochemistry results in samples obtained from a peripheral IVC to compare these results to those of samples obtained by direct venipuncture. Assessing the accuracy of samples drawn from a large gauge catheter is critical to proper patient care because clinical decisions are made based on the reliability of hematology results. The objective of this study was to determine the agreement between results for routinely measured hematologic analytes in blood collected from an IVC as compared with blood collected by direct venipuncture in hospitalized patients.
Materials and Methods
This study was conducted as a prospective, blinded, cross-over design study involving client-owned horses at the University of Pennsylvania New Bolton Center between August 2008 and June 2009. Horses were eligible for inclusion if they were ≥1 year of age, had a single jugular IVC placed for clinical therapy, and had a grossly normal contralateral jugular vein on palpation (ie, no signs of thrombosis, phlebitis, or thrombophlebitis). To evaluate the effect of fluid therapy on variables measured in blood drawn from the IVC, 50% of the study population was selected to consist of horses receiving constant rate infusion IV crystalloid fluids (on fluids), and 50% receiving only low volume IV medications such as antimicrobial or anti-inflammatory drugs (off fluids). To avoid type II error, power calculations (http://www.danielsoper.com/statcalc/calc47.aspx) were performed, and identified approximately 25 horses as an appropriate sample size for each arm of the study (50 horses total). Before the experiment, dead space of the catheter and extension set was calculated by a saline displacement technique ex vivo. All procedures were approved by the University of Pennsylvania Institutional Animal Care and Use Committee and consent was obtained from the owner of each participating animal.
Signalment, catheter size and manufacturer, date of catheter placement, current diagnosis, and IV fluid type and rate were recorded in all cases. For horses on fluids, crystalloid infusions were stopped 5 minutes before sampling. Direct venipuncture was performed 1st, with a 20 G 1 in. needle and 20-mL syringe to collect 12 mL of blood from the upper one third of the jugular vein while occluding the vessel distally; 8 mL was transferred immediately into an evacuated, heparinized blood collection tube,a and 3 mL into an evacuated ethylenediaminetetraacetic acid tube.b The IVC sample was then immediately collected after discarding a standardized presample (300% of catheter dead space). An identical syringe was used to withdraw 12 mL of blood from the catheter extension set, which was transferred into equivalent tubes. Unused blood was discarded. The jugular vein was not occluded when blood was collected through the IVC. All aliquots were identified by numbered codes so that laboratory personnel were blinded to the origin of the blood sample. A plasma biochemistry, CBC, PCV, and total protein concentration (by refractometry) and stall-side plasma glucose were performed by the New Bolton Center Clinical Pathology Laboratory. CBC variables included total white blood cell count, segmented neutrophils, lymphocytes, eosinophils, basophils, red blood cells, platelets, hemoglobin, and hematocrit. Biochemical analysis was performed on plasma and included the analytes glucose, creatinine, sodium, potassium, chloride, total carbon dioxide, total calcium, phosphorus, total protein, albumin, aspartate aminotransferase, creatine kinase, γ-glutamyl transpeptidase, and total bilirubin. All samples were analyzed by the same automated analyzers.c,d Glucose was measured by 2 different techniques, both as part of the standard laboratory plasma chemistry analysis, as well as with a commercial bedside glucometere (Table 1). Samples were maintained at room temperature between collection time and analysis, which was performed within 60 minutes of collection.
|Venipuncture Sample Mean (± SD)||IVC Sample Mean (± SD)||On-Fluids Mean Delta (± SD)||Off-Fluids Mean Delta (± SD)||All Samples Mean Delta (± SD)|
|Glucose (analyzer) (mg/dL)||114.2 (± 27.8)||114.7 (± 28.5)||−1.7 (± 4.3)||0.8 (± 3.6)||−0.5 (± 4.1)|
|Creatinine (mg/dL)||1.4 (± 0.3)||1.4 (± 0.3)||0.0 (± 0.1)||0.0 (± 0.1)||0.0 (± 0.0)|
|Sodium (mmol/L)||135 (± 3)||135 (± 5)||0 (± 1)||−1 (± 4)||0 (± 3)|
|Potassium (mmol/L)||3.53 (± 0.59)||3.63 (± 0.51)||−0.10 (± 0.29)||−0.11 (± 0.55)||−0.10 (± 0.43)|
|Chloride (mmol/L)||99 (± 2)||99 (± 3)||0 (± 1)||0 (± 1)||0 (± 1)|
|Total CO2 (mmol/L)||29.1 (± 2.7)||28.8 (± 2.6)||0.2 (± 0.9)||0.4 (± 1.0)||0.3 (± 0.9)|
|Total calcium (mg/dL)||10.89 (± 1.26)||10.90 (± 1.27)||−0.04 (± 0.15)||0.01 (± 0.20)||−0.01 (± 0.18)|
|Phosphorus (mg/dL)||3.35 (± 1.26)||3.33 (± 1.23)||0.01 (± 0.05)||0.03 (± 0.08)||0.02 (± 0.07)|
|Total protein (g/dL)||6.4 (± 1.1)||6.3 (± 1.1)||0.0 (± 0.1)||0.0 (± 0.2)||0.0 (± 0.2)|
|Albumin (g/dL)||2.74 (± 0.47)||2.73 (± 0.47)||0.00 (± 0.08)||0.00 (± 0.06)||0.00 (± 0.07)|
|AST (U/L)||884 (± 3403)||865 (± 3288)||36 (± 164)||3 (± 8)||20 (± 116)|
|CK (U/L)||8552 (± 57525)||8599 (± 57951)||−117 (± 601)||22 (± 60)||−47 (± 429)|
|GGT (U/L)||31 (± 16)||31 (± 16)||0 (± 2)||0 (± 2)||0 (± 2)|
|Total bilirubin (mg/dL)||3.0 (± 1.9)||3.0 (± 1.9)||0.0 (± 0.1)||0.0 (± 0.1)||0.0 (± 0.1)|
|WBC/μL||8.21 × 103 (± 3.28 × 103)||8.21 × 103 (± 3.40 × 103)||−0.13 × 103 (± 0.41 × 103)||0.036 × 103 (± 0.54 × 103)||−0.05 × 103 (± 0.48 × 103)|
|RBC/μL||7.39 × 106 (± 1.83 × 106)||7.42 × 106 (± 1.84 × 106)||−0.01 × 106 (± 0.23 × 106)||−0.05 × 106 (± 0.34 × 106)||−0.03 × 106 (± 0.29 × 106)|
|HGb (g/dL)||12.2 (± 2.7)||12.2 (± 2.8)||0.0 (± 0.5)||−0.0 (± 0.7)||0.0 (± 0.6)|
|HCT (%)||34.3 (± 7.8)||34.4 (± 7.7)||0.0 (± 1.1)||−0.2 (± 1.6)||0.0 (± 1.3)|
|Platelets/μL||1.43 × 105 (± 0.47 × 105)||1.45 × 105 (± 0.50 × 105)||−0.02 × 105 (± 0.12 × 105)||−0.01 × 105 (± 0.10 × 10)||−0.02 × 105 (± 0.12 × 105)|
|Neutrophils/μL||5743 (± 3827)||5695 (± 3799)||137 (± 790)||−40 (± 543)||48 (± 677)|
|Lymphocytes/μL||2024 (± 1188)||2065 (± 1249)||−110 (± 600)||28 (± 131)||−41 (± 435)|
|Monos/μL||494 (± 332)||475 (± 304)||7 (± 137)||31 (± 172)||19 (± 155)|
|Eosinophils/μL||106 (± 97)||109 (± 100)||1 (± 55)||−7 (± 62)||−3 (± 58)|
|Basophils/μL||36 (± 45)||38 (± 45)||−3 (± 15.95)||−2 (± 10)||−2 (± 13)|
|PCV (%)||34 (± 7)||34 (± 7)||0 (± 1)||0 (± 2)||0 (± 1)|
|Total solids (g/dL)||6.5 (± 1.0)||6.5 (± 1.0)||0.1 (± 0.1)||0.0 (± 0.2)||0.0 (± 0.2)|
|Glucose (glucometer) (mg/dL)||116 (± 30)||115 (± 29)||1 (± 8)||2 (± 7)||2 (± 8)|
Data were tabulated, and standard descriptive analysis performed. In each subject, the difference (delta) was calculated between analytes obtained from each site (venipuncture versus IVC). The mean of the delta values for each analyte also was calculated, and the mean deltas analyzed as a function of group to identify the effect of fluid administration on agreement between sites. The paired comparisons by group (“on fluids” versus “off fluids”) and by blood collection site (venipuncture versus IVC) were performed using Student's t-test with compensation for unequal variances (Satterwaite's test) between the 2 groups. A Bonferroni correction for a 5% 2-tailed hypothesis test would imply a P value < .0027 in this investigation.f
Between August 2008 and June 2009, 50 horses with IVCs were included in this study. They ranged from 1 to 28 years of age (median, 9.5 years). Breeds represented included Thoroughbred (24), Standardbred (4), Warmblood (8), Crossbred (3), Paint (2), Miniature Horse (2), Quarter Horse (1), American Saddlebred (1), Connemara (1), Icelandic (1), Lusitano (1), Morgan (1), and Pony of the Americas (1). The group included 19 mares, 29 geldings, and 6 stallions. The horses were admitted to the hospital for a variety of medical and surgical conditions, including disorders associated with the gastrointestinal tract (30), musculoskeletal system (8), respiratory system (3), ophthalmological abnormalities (4), or other disturbances (5). In the “on fluids” group, all horses were receiving a commercial isotonic crystalloid solution. The median fluid rate was 1.5 L/h (range, 0.15–3 L/h). Of the horses receiving fluids, 22/25 were supplemented with potassium, 5/25 with dextrose, and 4/25 with calcium gluconate. Where data were available (37/50 horses), the duration of placement of the IVC ranged from 2 to 336 hours. The types of catheters placed included 14 G × 20 cm single lumen polyurethane catheters (n = 19),g 14 G × 13.3 cm fluorinated ethylene-propylene catheters (n = 27),h,i and 14 G × 13.3 cm polyurethane catheters (4).j
One hundred samples were collected from 50 patients. The impact of catheter size on sample collection could not be analyzed because all adult patients had a 14 G catheter placed with a standard 7 in. large bore extension set without valves.k The effect of type of medication administered through the catheter was not assessed because of small sample size for each drug administered. There was no significant difference between venipuncture and IVC for any of the measured analytes (Table 1, left columns). There also was no significant difference between the mean delta values of any analyte in patients “on fluids” versus “off fluids” (Table 1, right columns).
Additionally, the agreement between samples drawn by venipuncture in comparison to those drawn directly from the IVC was subjectively determined to be clinically acceptable for all measured variables (Table 1). One patient had a significant effect on the standard deviation for mean creatine kinase and aspartate aminotransferase due to a diagnosis of acute rhabdomyolysis, but the delta between sites was minimal. In individual patients, where electrolytes or dextrose had been added to the IV fluids, the difference in the values of those specific analytes (potassium, calcium, and dextrose) subjectively was not clinically relevant between sampling sites. Although no statistically significant effect existed between the “on fluid” versus “off fluid” group, an occasional single analyte from the CBC had a larger delta value between sites. Hemodilution caused by crystalloid fluid administration could not be used to explain this effect because the abnormalities were sporadic within a patient. These results are reflected in the occasionally wide standard deviations of the delta values for analytes in the CBC.
We found no statistically significant difference in plasma biochemistry and CBC results from blood obtained by direct venipuncture versus blood acquired from a jugular catheter after removal of an adequate presample. Moreover, not only were the sample results not statistically different, but the differences were not felt to be large enough to affect clinical practice. Possible reasons for the small disparities in sample results observed in this study were real differences between the composition of blood in the different samples and bias associated with the precision limits of the measurement techniques used.
One of the most important concerns about blood collection by IVC is the increased occurrence of hemolysis, as evidenced either by gross plasma discoloration or by decreased hematocrit, mean corpuscular volume, sodium and glucose concentrations, and increased mean corpuscular hemoglobin, albumin, ionized calcium, potassium, and phosphorus concentrations.10 The main determinant of hemolysis appears to be small catheter internal diameter.4,8 One study suggests that hemolysis does not occur in catheters that are ≥16 G but does increase with a decreasing catheter diameter4 while another found mild hemolysis in half of the samples submitted for analysis after collection from 14 and 16 G catheters.8 Both studies were limited by small sample size with the former study drawing its conclusions on 3 and the latter on 4 samples. We did not detect any hemolysis in samples drawn from 14 G catheters that were ≤20 cm in length, but cannot draw conclusions about the potential for erythrocyte disruption in samples from smaller diameter catheters or those of greater length. Based on our data and previous studies,4 the risk of hemolysis with 16 G catheters seems low, suggesting that this technique is probably reliable in foals and other large animal species with indwelling catheters of this gauge.
Three accepted methods are used to remove residual fluid from a catheter before drawing a blood sample for analysis. These include the presample discard method (used in this study), the reinfusion method, where the heparinized presample is returned to the patient, and the push-pull method, where blood is aspirated and reinjected 3 times without disconnecting the syringe from the catheter. Conceptually, there may be an increased risk of introduction of contaminants or promotion of bacterial growth by withdrawing a blood sample through a catheter rather than using a catheter only for administration of drugs or fluids. Two observational studies in human medicine showed an increased risk of catheter-related bloodstream infection with excessive handling of the catheter.11 Additional investigation is warranted to establish if this technique is associated with increased complications (eg, phlebitis carries bacteremia) in hospitalized horses. Although data from humans would suggest this technique carries low risk, the more contaminated environment of a horse stall may pose additional risk in this population.
Coagulation studies were not performed on these samples, but small studies in animals and humans suggest that assessment of coagulation indices is reliable when samples are drawn through a central or peripheral catheter. We would expect similar results in the horse, where larger gauge catheters are more commonly used. A study in 35 dogs admitted to an ICU compared parameters (prothrombin time, activated partial thromboplastin time, and fibrinogen) in samples collected via catheter as compared with direct venipuncture.12 The gauge of the catheters was not discussed in the article but agreement between the 2 samples was found to be clinically acceptable. A similar test was carried out in healthy dogs and came to similar conclusions.9 Human studies demonstrate comparable results in patients with 18, 20, and 22 G peripheral venous catheters, in which aberrant findings would be anticipated because of the smaller gauges used in the study.1
In conclusion, blood samples obtained via IVC have clinically equivalent values to those taken by venipuncture in commonly performed analyses. The conclusions of this study are particularly robust in that we found no difference in analytes between sites in a clinical population of horses, as opposed to in experimental animals that may have fewer confounding influences. Sampling blood through a previously placed IVC is a viable option to decrease stress to the patient and increase positive interaction and ease of sampling for patient and staff. It should be stressed that deviation from the above-described protocol may lead to inconsistent results and that extrapolations by different methods of drawing blood through the catheter or when using catheters of different lengths and gauges may be inaccurate. Further studies should be performed to assess whether increased complications arise in horses with blood drawn repeatedly from the catheters.
aBD vacutainer lithium heparin (8 mL) tube, Fisher Scientific, Houston, TX
bBD vacutainer K2 EDTA 3.6 mg (3 mL) tube, Fisher Scientific
cVitros 350 Chemistry System, Ortho Clinical Diagnostics, Piscataway, NJ
dAdvia 120 Hematology System, Siemens, Deerfield, IL
eAccu-Chek Aviva, Roche, Nutley, NJ
fStata 10.0, StataCorp, College Station, TX
gCentral Venous Catheter, Arrow International, Reading, PA
hAngiocath, Deserte Medical Inc, Sandy, UT
iAbbocath-T, Abbott Laboratories, North Chicago, IL
jMilacath, Mila International, Covington, KY
kET-06L 7 inch extension set, International Win, Kennett Square, PA
The authors thank Sue Lindborg for all her technical assistance in performing this study. This study was supported by a grant from the Raymond Firestone Foundation (2008).
Conflict of interest: None.
- 11Assessment of infants with peripherally inserted central catheters: Part 1. Detecting the most Frequently Occurring Complications. Adv Neonatal Care 2002;2:304–315.