The accurate standardization of the pre-analytical phase is of pivotal importance for achieving accuracy and precision when measuring prothrombin time (PT), activated partial thromboplastin time (APTT), fibrinogen and D-dimer [1]. Among the major determinants of pre-analytical variability, unsuitable blood drawing techniques can influence the reliability of laboratory testing. Although venepuncture is traditionally carried out using ordinary straight needles, the butterfly device, a small needle attached to flexible plastic wings and connected with extension flexible tubing, might be regarded as a reliable alternative to collect blood in exceptional circumstances. In laboratory practice, the use of such a device has been historically advised for reasons of cost and when there is significant chance of obtaining unsuitable samples (i.e. uncompleted tube filling, hemolysis, activated samples). Therefore, to investigate the influence of a butterfly device on routine coagulation testing, blood was collected by a single expert phlebotomist into siliconized vacuum tubes, using either a 21G, 0.80 × 19 mm Venoject® multisample straight needle (Terumo Europe NV, Leuven, Belgium) (sample A), or a 21G needle butterfly device and 300 mm grade polyvinyl chloride (PVC) tubing with a Luer adapter (Artsana, Casnate, CO, Italy) (sample B). Venepunctures were performed in the morning of the same day on 30 fasted volunteers (18 women, 12 men; mean age 47 years); all phases of sample collection and preparation were standardized. No specimens were discarded for unsatisfactory attempts. PT, APTT and fibrinogen measurements were performed on the Behring Coagulation System (BCS, Dade-Behring, Marburg, Germany), employing proprietary reagents. Plasma D-dimer was measured using Vidas DD, a rapid and quantitative automated enzyme-linked immunosorbent assay, on the Mini Vidas Immunoanalyzer (bioMerieux, Marcy l'Etoile, France). All measurements were performed in duplicate within a single analytical session and final results were averaged. Analytical imprecision, expressed as a mean interassay coefficient of variation (CV), was quoted by the manufacturers as being between 2 and 5%.

The results are reported in Table 1. In each case, the means for paired samples collected by the two alternative drawing techniques did not differ significantly by paired Student's t-test. Bland & Altman plots and limits-of-agreement analysis showed mean biases between − 2.5% and 3.3% and relative CV ranging from 1.0% to 3.2%. The 95% agreement interval in the set of differences between values was acceptable and none of the tested parameters exceeded the current desirable analytical quality specifications for desirable total error [2], when these were available. Passing & Bablock regression analysis and relative correlation coefficients were satisfactory for all the analyses.

Table 1.  Statistical analysis of coagulation testing for specimens collected into evacuated tubes employing a 21G butterfly device and 300-mm PVC tubing (sample B) versus a 21G conventional straight needle (sample A)
 Sample ASample BPPassing & Bablock regression (r)CV (%)Desirable total error [2]Mean and (%) bias95% CI (%)
  1. Values are expressed as mean ± standard deviation. The difference between samples A and B is analyzed by paired Student's t-test (P) and by Passing & Bablock regression analysis and relative coefficient of correlation (r). The mean differences between samples A and B are shown as absolute and percentage bias, coefficient of variation (CV) and relative Altman & Bland 95% coefficient of interval limits of agreement (CI). Values are finally compared to the desirable analytical quality specifications for total error, as currently indicated by Ricos and colleagues [2].

PT (ratio)2.21 ± 0.732.20 ± 0.730.429y = 1.00x + 0.01 (r = 0.997)1.1± 5.3%− 0.007 (−0.3%)− 1.1 to 0.5
APTT (ratio)1.42 ± 0.281.42 ± 0.280.447y = 1.00x-0.01 (r = 0.998)1.0± 4.5%0.003 (0.2%)− 0.4 to 0.8
Fibrinogen (mg dL−1)395 ± 107397 ± 1110.481y = 0.96x + 15 (r = 0.996)1.4± 13.6%1.3 (0.3%)− 0.6 to 1.3
D-dimer (ng mL−1)546 ± 520548 ± 5340.777y = 0.99x + 3 (r = 0.997)3.22.2 (0.4%)− 2.5 to 3.3

Among laboratory tests, fibrinogen and D-dimer measurements are thought to be more susceptible to variations in the pre-analytical phase [3–5] and the choice of the device for drawing blood is of pivotal importance in achieving reliable results. As the surface hydrophobicity of most artificial surfaces induces hemostatic activation in vitro[6], it is conceivable that the blood flow within the 300-mm PVC tubing of the butterfly device might introduce variations from direct collection into vacuum tubes by traditional straight needles. PVC is extensively used for medical devices; however, it produces adverse reactions when in contact with body tissues and fluids and can lead to thrombus formation [7,8]. In particular, unmodified stents and membranes and non-coated circuits induce significant activation of coagulation [9,10] and a remarkable Factor XII-like activity is observed on the surface of the PVC [11]. Finally, the mechanical strain of the 300-mm-long tubing might affect the membrane integrity of blood leukocytes, erythrocytes and platelets, causing efflux of intracellular constituents and activating proteins. However, the present investigation indicates that there are no significant differences in the results of coagulation testing between blood specimens collected by either butterfly device or classical straight needle; in no case would a diagnosis have been failed or postponed, or an anticoagulant therapy substantially modified. The concordance of D-dimer measurements between samples A and B further suggests that, if an activation of the hemostatic system occurred within the PVC tubing of the butterfly device, this was modest and negligible in terms of laboratory testing. Therefore, we conclude that, when a proper technique is used and within certain limitations, the butterfly device may be a reliable alternative to the conventional straight needle for blood drawing for purposes of coagulation testing.


  1. Top of page
  2. References
  • 1
    Almagor M, Lavid-Levy O. Effects of blood-collection systems and tubes on hematologic, chemical, and coagulation tests and on plasma hemoglobin. Clin Chem 2001; 47: 7945.
  • 2
    Ricos C, Alvarez V, Cava F, Garcia-Lario JV, Hernandez A, Jimenez CV, Minchinela J, Perich C, Simon M. Current databases on biologic variation: pros, cons and progress. Scand J Clin Laboratory Invest 1999; 59: 491500.DOI: 10.1080/00365519950185229
  • 3
    Papp AC, Hatzakis H, Bracey A, Wu KK. ARIC Hemostasis Study I. Development of a blood collection and processing system suitable for multicenter hemostatic studies. Thromb Hemost 1989; 61: 1519.
  • 4
    Carroll PA, Ray MJ. Erroneous D-dimer result on a paediatric citrated specimen. Blood Coagul Fibrinolysis 1996; 7: 502.
  • 5
    Schutgens RE, Haas FJ, Ruven HJ, Spannagl M, Horn K, Biesma DH. No influence of heparin plasma and other (pre)analytic variables on D-dimer determinations. Clin Chem 2002; 48: 161113.
  • 6
    Hunt BJ, Parratt R, Cable M, Finch D, Yacoub M. Activation of coagulation and platelets is affected by the hydrophobicity of artificial surfaces. Blood Coagul Fibrinolysis 1997; 8: 22331.
  • 7
    Elam JH, Nygren H. Adsorption of coagulation proteins from whole blood on to polymer materials: relation to platelet activation. Biomaterials 1992; 13: 38.
  • 8
    Yianni JP. Making PVC more biocompatible. Med Device Technol 1995; 6: 206.
  • 9
    Urlesberger B, Zobel G, Rodl S, Dacar D, Friehs I, Leschnik B, Muntean W. Activation of the clotting system: heparin-coated versus non-coated systems for extracorporeal circulation. Int J Artif Organs 1997; 20: 70812.
  • 10
    Christensen K, Larsson R, Emanuelsson H, Elgue G, Larsson A. Coagulation and complement activation. Biomaterials 2001; 22: 34955.
  • 11
    Lamba NM, Courtney JM, Gaylor JD, Lowe GD. In vitro investigation of the blood response to medical grade PVC and the effect of heparin on the blood response. Biomaterials 2000; 21: 8996.