Administration of fibrinogen concentrate combined with prothrombin complex maintains hemostasis in children undergoing congenital heart repair (a long‐term propensity score‐matched study)

Abstract Background Bleeding is a common problem in children with congenital heart disease undergoing major cardiac surgery requiring cardiopulmonary bypass (CPB). Little is known about optimal management with blood products. Objective To investigate clinical outcome and hemostatic effects of fibrinogen concentrate (FC) in combination with prothrombin complex concentrate (PCC) versus standard treatment with fresh frozen plasma (FFP) in children undergoing cardiac surgery. Methods For this single‐institution cohort study, data on 525 children were analyzed. Propensity score matching in 210 children was applied to reduce the impact of various baseline characteristics. Results Three children treated with FC/PCC developed surgical site bleeding requiring surgical revision. One child developed central venous line‐related thrombosis. Blood loss through chest tube drainage was independent of FC/PCC. Coagulation abnormalities were not present in any of these children. Time to extubation and ICU stay did not differ. In the FC/PCC group, children received (median, Q1, Q3) 52 mg/kg (32, 83) FC and 28IU/kg (13, 44) PCC. Fibrinogen concentration was comparable at baseline. On admission to the ICU, fibrinogen was higher in children receiving FC/PCC, namely, 232 mg/dL (196, 280), than in children receiving FFP (186 mg/dL, 149, 224; P < .001). On discharge from the ICU, values did not differ ((FC/PCC 416 mg/dL (288, 501)), non‐FC/PCC 418 mg/dL (272, 585; P = 1.000)). Conclusion FC/PCC was well tolerated and permitted hemostasis to be maintained, even in the very young. We were not able to detect a signal for inferiority of this treatment. We conclude that FC/PCC can safely replace FFP.


| INTRODUC TI ON
Bleeding is a common problem in children undergoing major cardiac surgery requiring cardiopulmonary bypass (CPB). [1][2][3][4] Selection and quantity of blood products used in these severely ill children differ significantly between centers. [5][6][7] Information on optimal monitoring and management of hemostasis during congenital heart surgery in children is scarce in the literature. Encouraging outcomes in children and adults receiving fibrinogen led to integration of fibrinogen concentrates in the treatment algorithm at our institution. 6,[8][9][10][11][12][13] The aim of our study was to exploratively investigate the hemostatic effects of fibrinogen concentrate (FC) and prothrombin complex concentrate (PCC) treatment of children with congenital heart disease when undergoing heart surgery requiring CPB. Primary endpoint was the hemostatic effects of FC/PCC compared to treatment with FFP.
Data collection and analyses included bleeding, thrombosis, chest tube drainage (CTD), need for blood products, time to extubation, duration of ICU stay, fibrinogen levels, and PT/aPTT at selected time points.

| ROUTINE MANAG EMENT
Anesthesia was induced with midazolam, ketamine, fentanyl and rocuronium and maintained with remifentanil and sevoflurane. Oxygenator and circuit were primed with Ringer's lactate 3 ml/kg, mannitol 15% 3 ml/kg, 20 ml/kg human albumin (Albiomin ® 50 g/l; Biotest Pharma GmbH), and 10 mg/kg tranexamic acid. For children weighing less than 15 kg, the priming volume was composed of washed and filtered packed red blood cells (PRBC) and fresh frozen plasma (FFP, 20 ml/kg). The target was hematocrit between 28% and 30% and normal colloid osmotic pressure. Children with more than 15 kg received clear priming. Potassium, sodium, calcium, and magnesium were added to balance electrolytes. Before starting CPB, mucosa heparin (Baxter AG) was administered at 300 U/kg to achieve an activated clotting time (ACT) above 480 s. Additional doses of 100 U/kg were administered to maintain ACT in the target range.
Pump flow rates were maintained between 2.4 L and 3.0 L/ m 2 to ensure a mean target perfusion pressure between 40 and 50 mmHg (in neonates between 30 and 50 mmHg). All children independent of age routinely underwent hemofiltration during CPB. We used multiBic ® hemofiltration solution 4 mmol/L potassium (Fresenius Medical Care) to replace intravascular volume. We aimed to control intravascular volume and prevent hemodilution and edema.
Myocardial protection was achieved with intermittent antegrade potassium colloid cardioplegia. Tranexamic acid was administered as follows: 10

Editorial Comment
This single center retrospective study demonstrates that fibrinogen concentrate combined with prothrombin complex can be effectively used as an alternative to fresh frozen plasma in children undergoing congenital cardiac surgery.
FFP was reserved for children with continuously prolonged PT and aPTT, even after PCC administration.

| Demographic and clinical parameters
We recorded: age (months), body weight (bw, kg), gender, type of surgery, time on bypass (min), clamping time (min), time to extubation (hours), ICU stay (days), and chest tube drainage (CTD, ml/kg). CTD was measured every 24 h during ICU stay.
Any clinical complication of treatment was recorded.

| Blood sampling
Blood samples were obtained exclusively from venous lines. Baseline blood samples were obtained the day before surgery (TP 1). Another blood sample was taken on arrival at the ICU (TP 2), after 24 h at the ICU (TP 3), and before discharge from the ICU, but latest after 8 days at the ICU (TP 4). Blood cell counts were measured from 1.2 ml tubes containing 1.6 mg ethylenediaminetetraacetic acid (EDTA)/ ml blood (Sarstedt, Nuermbrecht, Germany). For coagulation tests, blood samples were collected in 1.8-ml tubes containing 0.3 ml (0.106 mol/l) buffered (pH 5.5) sodium citrate (Sarstedt). Chemical analysis was performed from 1.2 ml tubes containing 35 U/ml lithium heparin.

| Timing and Laboratory Analysis
Blood cell count was measured using the XE-5000 Analyzer

| S TATIS TIC AL ANALYS IS
Propensity score matching was used to reduce the impact of differ- (ml/kg), were entered into a logistic regression analysis with fibrinogen treatment as the dependent variable. FC was always combined with PCC, and therefore PCC was not included into the propensity score matching. From this model, a propensity score for each patient representing the probability of being treated with fibrinogen was derived as described by Austin. 14 The resulting propensity score was used to create a 1:1 matched pair patient subsample, using the nearest-neighbor matching method with a caliper width of 0.05.  Coagulation abnormalities were not present in any of these children.

Score-matched patient population
FC/PCC did not alter chest tube drainage.

| Blood cell count, coagulation parameters, and C-reactive protein in the Propensity Score-matched patient population
Blood cell count, coagulation parameters, and CRP measured at baseline (T1), on arrival at the ICU (T2), after 24 h at the ICU (T3), and immediately before discharge from the ICU (T4) are shown (Table 2).
At baseline, hemoglobin was comparable between groups. On admission to the ICU, after 24 h at the ICU and at discharge from the ICU, hemoglobin did not differ between groups. White blood cell count and platelet count did not differ between groups at any time point.  (34%) did not receive FC (Table S1). Children receiving FC were younger (P < .001), had a lower body weight (P < .001), longer time on cardiopulmonary bypass (P < .001), longer time to extubation (P < .001), and longer stay at the intensive care unit (ICU, P < .001). Amount of administered FC was 61 mg/kg (42, 98) vs none in the other group (P < .001). PCC was given in the FC group at 30 IU/kg (18,47) vs none in the other group (P < .001). Platelet concentrate requirements did not differ between groups. Chest tube drainage was similar between groups. Relative frequency of different surgery types significantly differed between children receiving FC and children not receiving FC (P = .001). Specifically, neonatal heart surgery was more common in children receiving FC (P = .002), while atrial/ventricular septal defect closure was more common in children not receiving FC (P = .001).

| Blood cell count, coagulation parameters, and C-reactive protein in the full unmatched population
Blood cell count, coagulation parameters, and CRP measured at baseline (T1), on arrival at the ICU (T2), after 24 hours at the ICU (T3), and immediately before discharge from the ICU (T4) are shown in the full unmatched population (Table S2). At baseline, hemoglobin was comparable between groups. On admission to the ICU, Hb was higher in the FC/PCC group (P < .001). After 24 h at the ICU, Hb was still higher in the FC/PCC group (P = .008). On discharge from the ICU, hemoglobin did not differ between groups. Although clinically not relevant, white blood cell count was higher in children receiving FC/PCC at baseline (P < .001). At all other time points, white blood cell count did not differ between groups.
Platelet count was comparable between groups at baseline, after admission to the ICU, after 24 h at the ICU, and at discharge from the ICU. CRP levels did not differ between age groups at baseline or on admission to the ICU.
After 24 h at the ICU and at discharge, CRP levels were higher in children receiving no FC/PCC.

| D ISCUSS I ON
Bleeding is a common problem in children undergoing major car-    there is a tendency to use more factor concentrates in this group of critically ill children. 16,17,20,21  We conclude that FC/PCC was well tolerated and permitted hemostasis to be restored and maintained and may be a safe alternative to FFP. 2,5,16,17,20,[22][23][24] Guidelines issued by the European Society of Anesthesiology underweight, and other peculiarities of these severely diseased children overruled age-specific aspects of the coagulation system.

| Limitations
The study design (cohort study) and the prolonged time interval of recruitment may have influenced results. However, all children were treated following strict in-house guidelines; any changes were pro- All children that received FC also received PCC. The group of children treated with FC/PCC was younger and differed in many aspects.
Propensity score matching was applied to accommodate this fact.
Another important potential confounder was priming of cardiopulmonary bypass systems with PRBC and FFP in younger children weighing <15 kg. 28 Again, propensity score matching was applied to reduce the risk of misinterpreting findings.

| CON CLUS ION
We compared FC/PCC and FFP in a large cohort of critically ill children undergoing major cardiac surgery requiring cardiopulmonary bypass. FC/PCC was well tolerated and permitted hemostasis to be maintained, even in the very young. Over the years FC/PCC treatment has been established at our institution with no perceived and recorded increase in adverse events including the need for blood products and reoperation. We conclude that FC/PCC allows hemostasis to be restored and maintained.

ACK N OWLED G M ENTS
Assistance with the article: VR, MM, WP and DW collected data; J.K:, DV, MM and DF coordinated collection and analysis of data; JF performed statistical analysis; WS, C.V-S. and HT wrote the manuscript; C.V-S. and WS led study implementation and analysis efforts; and all authors contributed to the design, analyzed data and edited the manuscript. The authors are indebted to Simon Reider for assistance in data retrieval and to Sigrid Egger and Anton Jeller for excellent technical assistance.

CO N FLI C T O F I NTE R E S T S
C.V-S. has received personal fees outside the submitted work from