Impact of blood sampling technique on reproducibility of viscoelastic coagulation monitor (VCM™) system test results in the neonate.

To evaluate the reproducibility of the results of the viscoelastic coagulation test (VCT) performed with a new viscoelastic coagulation monitor (VCM™ - Entegrion) on native blood obtained by heel prick blood sampling with two different techniques compared to the standard blood collection in the newborn. Three blood samples were tested with the VCM analyzer in each of the 67 study subjects admitted to our level 3 neonatal intensive care unit. Standard blood collection (S) was performed by direct puncture of a peripheral vessel or by drawing of blood in a syringe connected to an arterial or venous catheter. Then, two more blood samples were drawn through a single heel prick. The first heel prick blood sample (HP1) was collected in the sample well through the attached metal capillary while the second (HP2) was poured directly into the sample well. Blood samples were automatically drawn into their pre-warmed cartridges and inserted into the VCM analyzers set up for analyses, which ran for one hour. VCT blood variables included clotting time (CT), clot formation time (CFT), angle alpha (α), amplitude at 10 and 20 min (A10 and A20), maximum clot firmness (MCF), and lysis indexes at 30 and 45 min (LY30 - LY45). Agreement was quantified by calculating the mean difference and SD between measurements of VCT blood variables from S, HP1 and HP2 blood samples. The 95% limits of agreement were calculated by the Bland & Altman method, using the upper or lower limit of agreement to interpret the variability of the measurements. The Kendall's correlation coefficient evaluated the interdependence between SD and intra-measurement mean. S blood samples were easily obtained in all the study subjects, while mild difficulties were recorded in 3/67 infants (4.5%) with the HP1 blood sampling and in 5/67 infants (7%) with the HP2 blood sampling. Pairwise comparison of test results performed on blood samples drawn with HP1 and HP2 techniques showed moderate agreement for CT and α-angle, strong agreement for CFT, LY30 and LY45 and almost perfect agreement for A10, A20 and MCF. In pairwise comparison of VCM analyses performed on blood samples drawn with S technique vs HP1 and HP2 techniques, Kendall's correlation coefficient was significant for CT (S vs HP1 and HP1 vs HP2), CFT (S vs HP1 and S vs HP2), α-angle (S vs HP1) and MCF (S vs HP1). This suggests that the measurement error depends on the extent of the measurement. The overall ICC for blood sampling techniques ranged from 0.289 to 0.879 with best agreement observed for CFT (strong) and for A10, A20 and MCF (almost perfect). The LY30 index was the least repeatable measurement (poor agreement). The VCM analysis performed on the blood sample drawn with the HP1 technique showed the best repeatability compared with that performed with the S blood-sampling technique. VCT test results performed with the VCM analyzer on native blood drawn by heel prick in neonates are comparable to those obtained from standard blood samples. This could allow for a widespread, real-time assessment of the overall bedside haemostasis of these small patients.