RESUMO
Introduction: Blood samples having inappropriate volume are a substantial part of preanalytical errors. Inadequate sample volume for glycated haemoglobin (HbA1c) test may be a common problem of patients with diabetes mellitus having vascular changes. In this study, we compared HbA1c concentrations of underfilled and appropriately filled blood collection tubes. Materials and methods: To compare HbA1c concentrations, blood samples were collected into 2 mL tubes containing K3-EDTA from 109 subjects. Two blood samples (underfilled and appropriately filled) were drawn from a patient by the same personnel and materials. HbA1c measurements were assayed on a Cobas 6000 analyser module c 501 (Roche Diagnostics, Mannheim, Germany). The HbA1c% results were compared by t-test and Wilcoxon's signed-rank statistical methods (SPSS Inc., Chicago, USA). Bias analysis was performed using Microsoft Excel 4.0. Results: Underfilled samples were classified three groups (group 1, N = 44; group 2, N = 36; and group 3, N = 29) according to the filling ratio of the samples; 0.5 mL and below (< 25%), 0.5-1.0 mL (25-50%), and 1.0-2.0 mL (> 50%), respectively. When we compared underfilled tubes with pairing filled tubes, there was a statistically significant difference only with tubes filled less than 25% (P = 0.030). Furthermore, we have done bias analysis between paired tubes according to the diagnostic cut-off value of 6.5%. The bias was more prominent in up to 50% underfilled blood tubes (1.1%), when HbA1c concentrations were below the diagnostic cut-off of 6.5%. Conclusions: We suggest that the blood tubes with EDTA for HbA1c measurement should be filled with at least 50% to avoid clinical variations.