• Giuseppe Lippi, Gian Luca Salvagno, Martina Montagnana, Giorgio Brocco, and Gian Cesare Guidi.
• Istituto di Chimica e Microscopia Clinica, Dipartimento di Scienze Morfologico-Biomediche, Università degli Studi di Verona, Piazzale, Scuro, Verona, Italy. ulippi@tin.it
• Clin Chem Lab Med. 2006 Jan 1;44(8):1009-14.

BackgroundDespite remarkable advances in technology and laboratory automation, results of laboratory testing still suffer from a high degree of preanalytical variability. Although there is no definitive evidence, the use of small-gauge needles for venipuncture is usually discouraged to reduce the chance of producing unsuitable specimens.MethodsThe purpose of this investigation was to assess the influence of the needle size used to collect venous blood on the measurement of 14 common analytes, including free hemoglobin, the most representative enzymes, protein-bound substances and electrolytes. Results for venous blood samples collected from 20 fasting voluntary physicians using either a 23- (0.60 mmx19 mm) or 25-gauge-needle (0.50 mmx19 mm) butterfly devices with polyvinyl chloride tubing (1.40 mmx300 mm) were compared with reference specimens collected using a 21-gauge-needle (0.80 mmx19 mm) butterfly device with polyvinyl chloride tubing (1.40 mmx300 mm).ResultsAll means for paired samples collected using the smaller needles did not differ significantly from the reference specimen by paired Student's t-test analysis. Passing-Bablok regression analysis and Pearson's or Spearman (creatine kinase, aspartate aminotransferase, alanine aminotransferase and chloride) correlation were acceptable for most of the analyses, although a lower correlation coefficient was observed for electrolytes. In addition, when expressed as a percentage of the mean for paired samples, the s(y,x) value exceeded the desirable bias for free hemoglobin, glucose, lactate dehydrogenase, aspartate aminotransferase, sodium, chloride, calcium and magnesium (in samples collected using both 23 G and 25 G needles) and potassium (in samples collected using a 25 G needle). Although Bland-Altman plot analysis and +/-1.96 SD agreement intervals for the set of differences between values was acceptable overall, the bias was rather broad for free hemoglobin and several critical electrolytes (calcium, chloride, potassium, sodium), exceeding the respective limits for desirable bias.ConclusionsThe results of our investigation indicate that 23 G needles, if handled correctly, will not introduce any statistically or clinically significant error to the measurement results compared to a 21 G needle. For the 25 G needle, we observed increased variability for potassium compared to a 23 G needle. Small-bore needles of 25 G or less cannot be universally recommended when collecting venous blood for clinical chemistry testing and should be reserved for selected circumstances, such as in patients with problematical venous accesses and newborns. In such cases, however, the bias introduced by the use of smaller needles should always be taken into consideration when interpreting test results.

11 keywords...

hide…