To Characterize Standard Laboratory Coagulation Parameters and Plasma Thrombin Generation in Very Preterm Infants and to Investigate Their Relationship to Clinical Outcomes

Very premature infants are at risk of bleeding complications and are frequently given plasma because of a perception that coagulation test results are abnormal. However, “abnormal” clotting times may simply be due to physiological immaturity. We hypothesized that prospective characterization of coagulation tests alongside assessment of thrombin generation would provide information on overall haemostatic balance in this vulnerable patient cohort.

In a prospective observational study, blood was drawn into citrated tubes from cord blood of neonates <30/40 and subsequently on day 1, 3 and fortnightly until 30/40 corrected gestational age from non-heparinised lines. Exclusion criteria included antenatal intraventricular haemorrhage and lack of informed consent. Platelet poor plasma was obtained by centrifugation of whole blood. Prothrombin time (PT), activated partial thromboplastin time (APTT), fibrinogen, procoagulant and anticoagulant factor activity were measured and tissue factor (TF)-stimulated thrombin generation was characterized in a calibrated automated thrombography assay with a 1pM TF stimulus using Thrombinoscope™ software. Control plasma was obtained from cord blood of term neonates.

Between April 2013 and June 2014, 92 patients <30/40 were admitted. Six infants were excluded and 86 recruited. Mean gestational age and birth weight were 27 (23.7-29.9) weeks and 1019 (530-1590) g respectively. Median (5^th-95^thpercentile) Day 1 PT, APTT, and fibrinogen were 17.5 (12.9-26.7) s, 82.9 (53.4-139.8) s and 1.4 (0.7-3.7) g/L respectively which were significantly prolonged compared with published results of term infants; p<0.001. Serial analysis revealed correction of PT, APTT and fibrinogen with increasing postnatal age; p<0.001.

Further analysis of procoagulant and anticoagulant mechanisms was performed in a subset of infants (N=16) where sufficient plasma was available from umbilical cord samples. Despite apparent “prolongations” in clotting times, we found that preterm peak thrombin generated was similar to term infants (94.3 (33.4-149.8) nM vs. 93.5 (74.3-133.8) nM; p=0.94). Minor, non-significant differences in the area under the thombin generation curve (endogenous thrombin potential) were observed (88 (37.1-126.3) nm/min vs. 101.3 (76.8-139.7) nm/min; p=0.09). Alpha-2-macroglobulin (α2M)-bound thrombin cleaves the fluorogenic thrombin substrate utilized in this assay and contributes to measured thrombin generation. However, in keeping with published data in a more mature cohort, α2M activity was lower in preterm vs. term infants. Consequently, we speculate that true ETP in preterm infants may be higher, undermining the current perception that these infants are at risk of bleeding due to “abnormal” clotting times. In keeping with published data in a more mature cohort, mean activity of procoagulant factors II, VII, IX and X were reduced in preterm vs. term infants (0.33 (0.18-0.5) IU/ml vs. 0.45 (0.35-0.6) IU/ml, p=0.003; 0.38 (0.14-0.57) IU/ml vs. 0.42 (0.31-0.59) IU/ml, p=0.29; 0.19 (0.09-0.50) IU/ml vs. 0.28 (0.19-0.37) IU/ml, p=0.004 and 0.33 (0.13-0.52) IU/ml vs. 0.43 (0.32-0.58) IU/ml, p=0.02) respectively. Conversely, activity of anticoagulant factors Protein C (0.14U/ml (0.06-0.24) IU/ml vs. 0.27 (0.18-0.39) IU/ml; p=0.002), free protein S (0.39 (0.28-0.55) IU/ml vs. 0.47 (0.36-0.59) IU/ml; p=0.02), antithrombin (0.22 (0.06-0.36) IU/ml vs. 0.52 (0.38-0.69) IU/ml; p=0.0001) and tissue factor pathway inhibitor (7.7 (2.6-16.9) vs. 10 (4.2-15.6) ng/ml; p=0.14) was lower in preterm vs. term infants. No differences in attenuation in thrombin generation in response to the anticoagulant activated protein C were observed in preterm vs. term infants, p=0.92.

In conclusion, in the largest prospective study to date of very preterm infants, we describe typical ranges for widely available coagulation tests. We also demonstrate differences in both procoagulant and anticoagulant pathways, to which standard clotting tests may have limited sensitivity. Finally, we show for first time that thrombin generation is similar in very preterm and term infants. These findings call into question the current practice of exposing infants to plasma products based on standard laboratory parameters, a practice which is associated with enormous potential harm and is not evidence based. These findings have the potential to impact on neonatal practice worldwide.