Coagulation and Transfusion Medicine / GLASS VS PLASTIC IN HEMOSTASIS TESTING Comparison of Samples Obtained From 3.2% Sodium Citrate Glass and Two 3.2% Sodium Citrate Plastic Blood Collection Tubes Used in Coagulation Testing Robert C. Gosselin, CLS, 1 Kim Janatpour, MD, 1 Edward C. Larkin, MD, 1 Yanlap P. Lee, CLS, 1 and John T. Owings, MD 2 Key Words: Glass tubes; Plastic tubes; Coagulation testing Abstract We sought to compare coagulation test results obtained from patients using 2 plastic blood collection tubes and the traditional glass blood collection tube. Blood specimens were obtained from 241 patients in 3.2% buffered sodium citrate using standard glass tubes, in 3.2% buffered sodium citrate in plastic tubes, and in 3.2% sodium citrate sandwich tubes (plastic within plastic). All samples were obtained and processed contemporaneously and tested for prothrombin time (PT) and activated partial thromboplastin time (aptt). Residual plasma was frozen at 70 C for future testing, including fibrinogen, antithrombin, plasminogen, protein C and protein S (functional and antigenic), dilute Russell viper venom time (DRVVT), ristocetin cofactor, factor XIII, D dimer, anti-xa activity, and prothrombin fragment. Although paired t test analysis revealed statistically significant differences (P <.05) between glass and plastic for PT, aptt, fibrinogen, protein C (functional and antigenic), functional protein S, DRVVT and confirmation method, antithrombin, and factor XIII, these differences were not considered clinically significant. Traditionally, coagulation testing has been performed with specimens in 3.2% or 3.8% buffered sodium citrate in siliconized glass tubes. More recently, there has been increasing demand to collect blood in plastic tubes to avoid the inherent hazards associated with glass, including injuries due to broken glass and the risk of broken glass during centrifugation. Changes of blood collection tubes for coagulation testing always cause concern because of the potential for in vitro activation of the clotting cascade, with alteration of clotting times and various markers of clot activation and consumption. Therefore, we sought to compare 2 plastic blood collection tubes with traditional glass blood collection tubes to determine whether significant differences exist between glass and plastic blood collection tubes for routinely ordered coagulation studies. Materials and Methods After approval from our institutional review board, blood samples were obtained from a group of random outpatient and hospitalized patients using standard phlebotomy techniques. We prospectively obtained blood samples from 241 patients. For each patient, 3 additional tubes of blood were collected in a randomized manner: 3.2% buffered sodium citrate using standard glass Vacutainer tubes (Becton Dickinson, Franklin Lakes, NJ), 3.2% buffered sodium citrate in plastic Vacutainer tubes (Becton Dickinson), and 3.2% sodium citrate VACUETTE sandwich tubes (small 3.0-mL plastic tube within a larger -ml plastic tube; Grenier Bio-one, Monroe, NC). The outside of the Grenier tube is constructed of polyethylene terephthalate plastic, and the inner vacuum Downloaded from https://academic.oup.com/ajcp/article-abstract/122/6/843/1759247 Am J Clin Pathol 2004;122:843-848 843 843 843
Gosselin et al / GLASS VS PLASTIC IN HEMOSTASIS TESTING tube (hence sandwich ) containing the anticoagulant, is made of polypropylene plastic. Limited patient demographics, including sex, age, International Classification of Diseases, Ninth Revision codes, and medication history, were recorded. All blood samples were obtained and processed contemporaneously. Once processed, all samples were tested for prothrombin time (PT) using Innovin (Dade Behring, Deerfield, IL) and activated partial thromboplastin time (aptt) using Actin FS (Dade Behring). Residual plasma was frozen at 70 C for later testing Table 1, including Dade Behring methods for clottable fibrinogen (MultiFibrin U), chromogenic antithrombin activity (Berichrom), chromogenic plasminogen activity (Berichrom), clottable protein C activity, dilute Russell viper venom time (DRVVT) and confirmation method or DRVVT ratio (LA1 and LA2, respectively), ristocetin cofactor (BC von Willebrand reagent), chromogenic factor XIII (Berichrom), D dimer (Advanced D-Dimer), anti-xa activity (Berichrom), and prothrombin fragment (F1.2) using an enzyme-linked immunosorbent assay (ELISA) method (Enzygnost). Functional protein S was tested using StaClot Protein S (Diagnostica Stago, Parsippany, NJ). Antigenic protein C and protein S were tested using ELISA methods (Helena Laboratories, Beaumont, TX). Except for the ELISA methods, all coagulation testing was performed on the BCS analyzer (Dade Behring), and each test was measured concurrently on the plasma collected from the 3 blood collection tubes. Owing to increased costs associated with special coagulation testing, only selected patient samples were tested, which included normal samples and samples from patients with known pharmaceutical anticoagulation (oral vitamin K antagonist or heparin), patients after trauma or surgery, and hospitalized medical patients. The selection of patient samples for special coagulation testing was based on the potential of obtaining normal and abnormal (less or more than normal reference intervals) results for a given test. The Pearson correlation coefficient, Student paired t test, and repeated measures analysis of variance were used to compare results between collection tubes. A P value of less than.05 was considered statistically significant. Modified Bland- Altman style bias plots were generated to ascertain areas of bias between blood collection methods. Results All tests performed on samples obtained in plastic tubes correlated significantly with results obtained from glass tube samples (R > 0.92; P <.05). Paired t test analysis revealed statistically significant differences between blood collection methods and laboratory parameters Table 2. There were no statistically significant differences (P >.05) between glass and BD plastic or Grenier plastic tubes for D dimer, plasminogen activity, ristocetin cofactor, anti-xa activity, prothrombin fragment, or protein S antigen. There were statistically significant differences (P <.05) between blood collection tubes for PT, fibrinogen, and functional protein S, in addition with statistically significant differences between glass and both plastic collection methods for DRVVT, antithrombin, and factor XIII activity. In addition, there were significant differences (P <.05) between glass and BD plastic for aptt, protein C activity and antigen, and DRVVT confirmation ratio. There was no statistically significant difference between glass and Grenier plastic tubes for aptt results. There were no statistically significant differences between BD plastic and Grenier plastic tubes for D dimer, protein C function and antigen, DRVVT, DRVVT ratio, plasminogen, factor XIII, anti-xa activity, prothrombin fragment, or protein S antigen. Table 1 Reagent Description, Manufacturer, and Test Method Principle Test Manufacturer Test Method Principle PT Dade Behring, Deerfield, IL Recombinant tissue factor activation aptt Dade Behring Purified soy phosphatides with ellagic acid Fibrinogen Dade Behring Clotting: clottable fibrinogen D dimer Dade Behring Latex turbidimetric: using DD5 antibody Protein C activity Dade Behring Clotting: modified aptt after snake activation Protein C antigen Helena Laboratories, Beaumont, TX ELISA: antihuman protein C capture antibody Protein S activity Diagnostica Stago, Parsippany, NJ Clotting: modified aptt using activated protein C and factor Va reagents Protein S antigen Helena Laboratories ELISA: antihuman protein S capture antibody DRVVT and DRVVT ratio Dade Behring Clotting: Russell viper venom activation of factor X Antithrombin Dade Behring Chromogenic: thrombin substrate Plasminogen Dade Behring Chromogenic: streptokinase activation Factor XIII Dade Behring Chromogenic: NAD-NADH measurements Ristocetin cofactor Dade Behring Platelet agglutination method Anti-Xa Dade Behring Chromogenic: dextran sulfate and factor Xa Prothrombin fragment Dade Behring ELISA: rabbit antihuman F1.2 capture antibodies aptt, activated partial thromboplastin time; DD5, D dimer 5; DRVVT, dilute Russell viper venom time; ELISA, enzyme-linked immunosorbent assay; NAD, nicotinamide adenine dinucleotide; NADH, reduced form of NAD; PT, prothrombin time. 844 Am J Clin Pathol 2004;122:843-848 Downloaded 844 from https://academic.oup.com/ajcp/article-abstract/122/6/843/1759247
Coagulation and Transfusion Medicine / ORIGINAL ARTICLE Table 2 Data Analysis Comparing Coagulation Test Results Obtained From Glass and Plastic Collection Tubes * 3.2% BD Glass 3.2% BD Plastic 3.2% Grenier Plastic P P P PT (s) (n = 241) 9.3 (7.4-56.1) 9.1 (7.3-54.4) 9.2 (7.6-54.7) <.001 <.001 <.001 aptt (s) (n = 241) 29.5 (18.1-144.9) 28.7 (17.6-117.8) 29.4 (18.1-134.6) <.001.78 <.001 Fibrinogen (mg/dl) (n = 112) 482.5 (226.5-948.3) 446.4 (23-922.2) 458.7 (230.8-1018.0) <.001.015.003 D dimer (mg/l) (n = 54) 2.50 (0.4-20.9) 2.35 (0.4-20.8) 2.60 (0.4-21.2).68.16.14 Protein C (%NHP) (n = 51) 97.8 (11.0-199.6) 97.0 (11.0-199.4) 97.6 (11.0-201.7).021.66.17 Protein S (%NHP) (n = 46) 100.9 (14.0-168.5) 108.4 (14.0-166.0) 99.0 (14.0-210.8).038 <.001.012 DRVVT (s) (n = 57) 38.8 (28.9-128.7) 40.3 (30.2-137.0) 41.9 (31.9-135.1) <.001 <.001.09 DRVVT confirmation ratio (n = 51) 1.20 (1.05-1.88) 1.21 (1.06-1.91) 1.22 (1.08-1.98).022.11.72 Antithrombin (%NHP) (n = 73) 91.4 (17.8-135.7) 89.8 (17.4-130.9) 89.2 (18.9-135.1).004 <.001.051 Plasminogen (%NHP) (n = 43) 112.6 (32.0-220.9) 111.5 (32.9-211.1) 113.7 (31.5-224.3).44.58.92 Factor XIII (%NHP) (n = 30) 99.3 (35.2-161.0) 99.5 (30.1-161.0) 99.7 (37.3-171.0).014.013.77 Ristocetin cofactor (%NHP) (n = 36) 102.6 (41.6-317) 98.5 (45.8-295) 104.0 (45.4-305).50.10.005 Anti-Xa activity (U/mL) (n = 44) 0.24 (-1.29) 0.22 (-1.10) 0.21 (-1.29).06.10.50 Protein C antigen (%NHP) (n = 33) 94.4 (37.3-132.0) 90.5 (33.5-128.1) 89.1 (37.6-148.0) <.001.09.93 Protein S antigen (%NHP) (n = 25) 106.6 (46.8-150.5) 98.4 (49.9-143.9) 98.3 (52.8-143.3).72.70.92 Prothrombin fragment (nmol/l) 0.89 (8-3.3) 0.80 (8-2.93) 0.87 (7-2.91).31.91.49 (n = 34) aptt, activated partial thromboplastin time; NHP, normal human plasma; PT, prothrombin time; DRVVT, dilute Russell viper venom time. * Data are given as median (range). P values were determined by using the Student paired t test. For proprietary information, see the text. Comparison between BD glass and BD plastic tubes. Comparison between BD glass and Grenier Bio-one tubes. Comparison between BD plastic and Grenier Bio-one plastic tubes. By using a modified Bland-Altman style bias plot, we demonstrated that the majority of PT values were within 2.0 seconds Figure 1A. For aptts less than 50 seconds, the majority of samples demonstrated a less than -second bias Figure 1B. The spread of aptt differences became larger at more than 50 seconds. There was a general trend of increasing bias with increased levels of protein and peptide and increasing clotting times Figure 1C, Figure 1D, Figure 1E, Figure 1F, Figure 1G, Figure 1H, Figure 1I, Figure 1J, and Figure 1K. We compared anti-xa activities with corresponding aptts in a small population of patients receiving heparin therapy (n = 15). By using regression analysis of these data, we calculated the aptt range based on 0.3 to 0.7 U/mL of anti-xa activity of unfractionated heparin to be 50 to 65 seconds, 48 to 59 seconds, and 50 to 63 seconds for glass, BD plastic, and Grenier plastic tubes, respectively. Four patients enrolled in the study were receiving direct thrombin inhibitors (DTIs) at the time blood samples were obtained: lepirudin, 2 patients; bivalirudin, 1 patient; and argatroban, 1 patient. In these DTI samples, the aptt values obtained from Grenier plastic tubes had a higher bias compared with samples obtained in glass tubes. To determine whether this was an aberrant finding, we obtained a blood sample from a healthy person using prepared lepirudin syringes containing between 0.1 and 1.2 µg/ml of the drug. Once drawn, the whole blood was dispensed into the 3 blood collection tubes (glass, BD plastic, Grenier plastic) and processed as previously described. These samples were analyzed for aptt, which also demonstrated a positive bias with the Grenier plastic tubes Figure 2. Discussion We sought to determine whether conversion to plastic sodium citrate tubes for coagulation testing would result in significant differences in results. There have been a limited number of reports on the impact of plastic blood collection tubes on coagulation testing, which are primarily limited to PT testing. 1-3 Another report indicated statistically significant differences in stratified patient populations for PT, aptt, fibrinogen, prothrombin fragment, and activated factor XII but not for anti-xa activity using Terumo Medical (Somerset, NJ) glass and plastic blood collection systems. 4 More recently, Flanders and colleagues 5 reported a statistical difference in the thrombin time test but not in other clotting tests in a small series of healthy patients. In the present study, we compared coagulation testing using traditional 3.2% sodium citrate glass tubes and 2 available 3.2% sodium citrate plastic tubes, using 5 testing methods (clotting, chromogenic, immunoturbidimetric, ELISA, and aggregation methods) on 16 analytes. Because all samples were obtained, processed, and tested contemporaneously for any given test, any bias noted would have been a result of the collection method. We demonstrated that a number of results reported from samples collected in traditional glass tubes were significantly different from results obtained from samples collected in plastic collection tubes. Most test biases tended to increase with increasing clotting times or protein levels. The majority of the biases seen with PT were less than 2.0 seconds. The Grenier plastic tube samples tended to give higher aptt results than did the glass tubes, whereas results with the BD plastic tubes were lower than with the glass tubes. In general, the discrepancies in the absolute values for any analyte obtained from Downloaded from https://academic.oup.com/ajcp/article-abstract/122/6/843/1759247 Am J Clin Pathol 2004;122:843-848 845 845 845
Gosselin et al / GLASS VS PLASTIC IN HEMOSTASIS TESTING A 6.0 4.0 2.0 4.0 B 2 1 1 2 C 2.0 1 2 3 4 5 6 7 Glass PT, s 3 5 10 15 20 Glass aptt, s D 10 5 5 10 15 20 150 350 550 750 950 1,150 Glass Fibrinogen, mg/dl Difference, mg/dl 15 2 1 1 1 0 25 50 75 100 125 150 Glass Antithrombin, %NHP E F 4 3 2 1 1 2 3 0 25 50 75 100 125 150 175 200 225 Glass Protein C, %NHP 5 4 3 2 1 1 2 0 25 50 75 100 125 150 175 200 Glass Protein S, %NHP G H 2 1 1 1 2 2 5 7 10 12 15 Glass DRVVT, s Difference, Ratio 0.30 0.20 0.10 0 0.10 0.20 0.30 1.00 1.25 1.50 1.75 2.00 Glass DRVVT Ratio 846 Am J Clin Pathol 2004;122:843-848 Downloaded 846 from https://academic.oup.com/ajcp/article-abstract/122/6/843/1759247
Coagulation and Transfusion Medicine / ORIGINAL ARTICLE I 40 30 20 10 0 10 20 30 J 80 60 40 20 0 20 0 25 50 75 100 125 150 Glass Protein C Antigen, %NHP 40 0 50 100 150 200 250 300 350 Glass Ristocetin Cofactor, %NHP K 1 1 1 0 25 50 75 100 125 150 175 Glass Factor XIII, %NHP Figure 1 Modified Bland-Altman style bias plots comparing the observed difference between results of samples collected in 3.2% sodium citrate glass tubes with results from samples collected in 3.2% sodium citrate plastic tubes from Becton Dickinson (squares) and Grenier Bio-one (triangles). Only tests that yielded significant differences between glass and plastic tubes (Table 2 data) were plotted. Data are given in conventional units for fibrinogen and factor XIII; to convert to Système International units for fibrinogen (g/l), divide by 1; for factor XIII (proportion of 1.0), divide by 1. aptt, activated partial thromboplastin time; DRVVT, dilute Russell viper venom time; NHP, normal human plasma; PT, prothrombin time. For proprietary information, see the text. any blood collection method would not have resulted in a change in clinical decision or treatment and, therefore, were deemed clinically insignificant. The exceptions are the apparent biases noted with samples obtained in Grenier plastic tubes from patients receiving DTI therapy and the limitations to be described later. There was no effect of blood collection method on the immunoturbidimetric method of analysis (D dimer). There was no effect of blood collection tubes on chromogenic determination of heparin and plasminogen, but blood collection tubes did affect antithrombin and factor XIII activity testing, with the majority of results within 5% of normal human plasma. It is unclear why the antithrombin test, using thrombin substrate, or factor XIII activity would be affected by plastic blood collection methods, although it has been reported that a thrombin-based test (the thrombin time) was significantly dependent on collection method. 5 Because thrombin is a component in the antithrombin and factor XIII activity measurements used in the present study, there might be some effect in the thrombin generation pathway that might be affected by the type of collection tube. There are several limitations to this study. First, although all samples were obtained concurrently and improperly obtained samples (clotted and/or incomplete filling) ultimately 1 1 2 2 5 7 10 12 Glass aptt, s Figure 2 Modified Bland-Altman style bias plots comparing the observed difference between results of samples collected in 3.2% sodium citrate glass tubes with results from samples collected in 3.2% sodium citrate plastic tubes from Becton Dickinson (squares) and Grenier Bio-one (triangles) in samples containing direct thrombin inhibitors (lepirudin, bivalirudin, or argatroban). aptt, activated partial thromboplastin time. For proprietary information, see the text. rejected, all blood samples were obtained using individual vacuum tubes. Therefore, subtle differences in tube blood volume, which would mimic clinical practice, might result in minor, clinically insignificant differences in the results. Downloaded from https://academic.oup.com/ajcp/article-abstract/122/6/843/1759247 Am J Clin Pathol 2004;122:843-848 847 847 847
Gosselin et al / GLASS VS PLASTIC IN HEMOSTASIS TESTING Second, when comparing samples from patients receiving heparin treatment, there were minor differences found in results from the 3 collection tubes. However, a limited number of heparin samples were tested (n = 15), and, therefore, further testing on samples obtained from patients receiving therapeutic heparin must be performed using the procedure described by Brill-Edwards and colleagues. 6 There also were a limited number of samples from patients receiving DTIs, and there seems to be a bias associated with Grenier plastic tubes. We tested samples obtained by in vitro anticoagulation before placement into blood collection tubes, but this might not accurately reflect in vivo drug response. In addition, because we tested any given patient only once during the duration of the study, it is unclear whether aptt ratios, which are the recommended target for DTI anticoagulation, would mitigate any collection bias seen with aptt testing. Conclusions There are statistically significant differences noted for coagulation testing of samples obtained in glass and plastic tubes. However, these differences are not clinically significant. As such, replacement of glass sodium citrate blood collection tubes with plastic sodium citrate blood collection tubes would be acceptable, with the following concerns: (1) We found some evidence that samples from patients treated with heparin or DTIs might be affected by collection method, and further evaluation in this area is necessary. (2) Because of the trending bias between plastic and blood collection tubes with increasing protein levels or clotting times, an adjustment to normal reference intervals might be warranted after reevaluating normal samples obtained in plastic tubes. From the Departments of 1 Pathology and 2 Surgery, University of California, Davis Medical Center, Sacramento. Address reprint requests to Dr Owings: UCDMC Trauma Division, Room 4209, 2315 Stockton Blvd, Sacramento, CA 95817. None of the investigators in this study has any conflict of interest with reagent or drug manufacturers. References 1. Tripodi A, Chantarangkul V, Bressi C, et al. How to evaluate the influence of blood collection systems on the international sensitivity index: protocol applied to two new evacuated tubes and eight coagulometer/thromboplastin combinations. Thromb Res. 2002;108:85-89. 2. Ridyard J, Bhavnani M, Seal LH. Laboratory control of oral anticoagulant therapy: preservation of prothrombin time specimens using a polypropylene collection system. Clin Lab Haematol. 1998;20:369-372. 3. D Angelo G, Villa C. Measurement of prothrombin time in patients on oral anticoagulant therapy: effect of two different evacuated tubes [letter]. Haematologica. 1999;84:656-657. 4. Biron-Andreani C, Mallol C, Seguret F, et al. Plastic versus siliconized glass tubes: evaluation in current laboratory practice [letter]. Thromb Haemost. 2000;83:800-801. 5. Flanders MM, Crist R, Rodgers GM. A comparison of blood collection in glass versus plastic Vacutainers on results of esoteric coagulation assays. Lab Med. 2003;34:732-735. 6. Brill-Edwards P, Ginsberg JS, Johnston M, et al. Establishing a therapeutic range for heparin therapy. Ann Intern Med. 1993;119:104-109. 848 Am J Clin Pathol 2004;122:843-848 Downloaded 848 from https://academic.oup.com/ajcp/article-abstract/122/6/843/1759247