- Email: [email protected]
Low Preoperative Fibrinogen Plasma Concentration Is Associated With Excessive Bleeding After Cardiac Operations
Low Preoperative Fibrinogen Plasma Concentration Is Associated With Excessive Bleeding After Cardiac Operations Katarina Wald en, MD, Anders Jeppsson, MD, PhD, Salmir Nasic, MS, Erika Backlund, RN, and Martin Karlsson, MD, PhD Department of Cardiothoracic Surgery, Sahlgrenska University Hospital, Gothenburg; Department of Molecular and Clinical Medicine, Institute of Medicine, Sahlgrenska Academy, University of Gothenburg, Gothenburg; Research and Development Centre, Skaraborg Hospital, Sk€ ovde; and Department of Medicine, Skaraborg Hospital, Lidk€ oping, Sweden
Background. Data from small selected patient populations suggest that the preoperative plasma concentration of fibrinogen influences postoperative blood loss and red blood cell transfusion after cardiac operations, but there are also conflicting reports. We assessed the importance of preoperative fibrinogen concentration for excessive bleeding and red cell blood transfusion in a large cohort of mixed cardiac surgical patients. Methods. We included 1,954 cardiac surgical patients in a prospective observational study. The fibrinogen plasma concentration was measured on the day before the operation. Blood loss (mediastinal drain volume) during the first 12 postoperative hours and red blood cell transfusion during the hospital stay were registered and related to fibrinogen concentration with logistic regression models. Excessive bleeding was defined as postoperative blood loss exceeding 1,000 mL/12 hours. Results. The preoperative fibrinogen concentration was inversely proportional to the prevalence of excessive
bleeding in univariate testing (odds ratio [OR], 0.75; 95% confidence interval [CI], 0.64 to 0.89 per g/L; p [ 0.001) and also in a multiple model adjusted for age, sex, body mass index, renal function, acuteness of the operation, cardiopulmonary bypass time, clopidogrel use less than 5 days before the operation, and type of operation (OR for fibrinogen, 0.82; 95% CI, 0.69 to 0.97; p [ 0.024). In contrast, the prevalence of red cell blood transfusion increased with increasing fibrinogen levels in univariate testing (OR, 1.36; 95% CI, 1.24 to 1.49; p < 0.001) but not in a multiple model (OR, 1.10; 95% CI, 0.89 to 1.28; p [ 0.49). Conclusions. Preoperative plasma concentration of fibrinogen is independently associated with excessive bleeding after cardiac operations but not with red blood cell transfusion.
F
Patients and Methods
ibrinogen is a key factor in the coagulation cascade. There are conflicting data on whether the preoperative plasma concentration of fibrinogen is associated with postoperative bleeding volume and transfusion after cardiac operations. Although some studies have found a statistically significant correlation with bleeding [1–7], contradictory reports exist [8–11]. Most of the previous studies have been too small to allow adjustment for confounding factors. In addition, most of them have included only elective coronary artery bypass grafting (CABG) patients. We sought to determine the relationship between the preoperative plasma fibrinogen concentration, on the one hand, and postoperative bleeding and red blood cell (RBC) transfusions, on the other, in a large unselected cohort of mixed cardiac surgical patients. Accepted for publication Nov 25, 2013. Address correspondence to Dr Jeppsson, Department of Cardiothoracic Surgery, Sahlgrenska University Hospital, 413 45 Gothenburg, Sweden; e-mail: [email protected].
Ó 2014 by The Society of Thoracic Surgeons Published by Elsevier Inc
(Ann Thorac Surg 2014;97:1199–206) Ó 2014 by The Society of Thoracic Surgeons
This study was approved by the Regional Research Ethics Committee, which waived the requirement for individual patient consent.
Patients From February 2009 to January 2011, 2,208 adult cardiac surgical procedures were performed at our institution. We excluded 170 (6.1%) procedures due to missing data, resulting in a study cohort of 1,954 procedures. All types of acute and elective procedures were included: CABG, valve operations, combined CABG and valve operations, adult congenital operations, and arrhythmia operations. Patients who underwent operations on the ascending aorta, including dissections and aneurysms, were excluded because of their high risk of surgical bleeding. If a patient underwent more than one operation during a single hospital admission, only the first operation was included in the analysis. A separate analysis was performed in isolated CABG patients to allow comparisons 0003-4975/$36.00 http://dx.doi.org/10.1016/j.athoracsur.2013.11.064
ADULT CARDIAC
CARDIOTHORACIC ANESTHESIOLOGY: The Annals of Thoracic Surgery CME Program is located online at http://www.annalsthoracicsurgery.org/cme/ home. To take the CME activity related to this article, you must have either an STS member or an individual non-member subscription to the journal.
1200
ET AL WALDEN FIBRINOGEN AND POSTOPERATIVE BLEEDING
Ann Thorac Surg 2014;97:1199–206
ADULT CARDIAC
with previous studies. Patient characteristics are given in Table 1. To further illustrate the importance of fibrinogen concentration, the patients were arbitrarily divided into five groups according to the preoperative fibrinogen concentration (2.5, 2.6 to 3.0, 3.1 to 3.8, 3.9 to 4.5, and 4.6 g/L). Patient characteristics for each group are reported in Table 2.
Laboratory Variables Blood samples were collected on the day before the operation (elective cases) or immediately before the operation (acute cases). Plasma fibrinogen concentration was determined according to the method of Clauss [12] (STA-FIB 2; Diagnostica Stago, Asnieres, France). The reference value is 2.0 to 4.5 g/L. Activated partial thromboplastin time, prothrombin time, serum creatinine, and hemoglobin were measured preoperatively with standard clinical methods. Estimated glomerular filtration rate was calculated with the Cockcroft-Gault formula [13].
Clinical Management Anesthesia in all patients was induced with fentanyl, thiopentone, and pancuronium and maintained with sevoflurane or propofol. The patients received heparin (300 U/kg body weight) to maintain an activated clotting
time exceeding 480 seconds. After cardiopulmonary bypass, the heparin was reversed by the administration of 1 mg protamine/100 U heparin. A cell-saving device was not used routinely. Cardioprotection was achieved with antegrade cold blood cardioplegia. Thromboelastometry/ graphy was not used routinely. In patients with on-going bleeding, thromboelastometry was used to assess coagulation and guide transfusions of blood products. Venous reservoir blood was not processed in a cell-saving device before retransfusion. Aspirin was not discontinued preoperatively. In elective cases, clopidogrel was discontinued 5 days before surgery. All patients received 2 g tranexamic acid (Cyklokapron; Pfizer Inc, New York, NY) preoperatively and 2 g postoperatively after weaning off cardiopulmonary bypass. Aprotinin was not used.
Bleeding and Transfusions Postoperative bleeding was defined as the total amount of chest tube drainage after closure of the sternum and during the first 12 postoperative hours, or until the patient was reexplored because of bleeding. The total amount of transfused packed RBCs (PRBCs), fresh frozen plasma, and platelets during the hospital stay was recorded. The local protocol specifies that PRBC transfusions should be given when the blood hemoglobin level has decreased to below 70 g/L or when a clinically
Table 1. Characteristics for All Patients and for the Subgroups With and Without Coronary Artery Bypass Grafting Variablea Age, y Female gender BMI, kg/m2 Smoking Diabetes Preoperative medication Aspirin Clopidogrel (<5 days pre-op) Fibrinogen, g/L aPTT, s PT (INR) Platelet count, 109/L Plasma creatinine, mmol/L eGFR, mL/min Hemoglobin, g/L Operative procedure CABG Aortic valve Mitral valve CABG þ valve Other Acute operation ECC time, min a
All Patients (N ¼ 1,954)
CABG Patients (n ¼ 1,075)
Non-CABG Patients (n ¼ 879)
p Value (CABG vs non-CABG)
66 12 489 (25) 27 4 228 (12) 402 (21)
67 9 235 (22) 27 4 154 (14) 278 (26)
64 14 254 (29) 26 4 74 (8) 124 (14)
<0.001 <0.001 <0.001 <0.001 <0.001
1,362 (70) 183 (9) 3.8 1.0 36 9 1.1 0.3 260 78 89 46 88 31 137 15
1,054 (98) 176 (16) 3.9 1.0 36 10 1.1 0.2 267 75 89 40 88 30 138 14
308 (35) 7 (1) 3.7 1.0 36 7 1.2 0.3 252 81 90 54 88 34 136 16
<0.001 <0.001 <0.001 0.69 <0.001 <0.001 0.54 0.82 0.004
1,075 (55) 387 (20) 154 (8) 205 (10) 133 (7) 88 (4.5) 95 42
1,075 – – – – 66 (6.1) 77 27
– 387 (44) 154 (18) 205 (23) 133 (15) 22 (2.5) 117 46
– – – – – <0.001 <0.001
Continuous data are shown as mean standard deviation and categoric data as number (%).
aPTT ¼ activated partial thromboplastin time; BMI ¼ body mass index; CABG ¼ coronary artery bypass grafting; ECC ¼ extracorporeal circulation; eGFR ¼ estimated glomerular filtration rate; INR ¼ internationalized normalized ratio; PT ¼ prothrombin time.
ET AL WALDEN FIBRINOGEN AND POSTOPERATIVE BLEEDING
Ann Thorac Surg 2014;97:1199–206
1201
Table 2. Characteristics Related to Preoperative Fibrinogen Concentration in All Patients (N ¼ 1,954)
Variable
Age, y Female gender BMI, kg/m2 Weight, kg CABG Diabetes ECC time, min Smoking Acute operation Aspirin Clopidogrel (<5 days pre-op) APTT, s PT (INR) Platelet count, 109/L Serum creatinine, mmol/L eGFR, mL/min Hemoglobin, g/L
2.5 (n ¼ 105)
2.6–3.0 (n ¼ 340)
3.1–3.8 (n ¼ 729)
3.9–4.5 (n ¼ 410)
4.6 (n ¼ 370)
57 18 16 (15) 25 4 76 13 29 (28) 9 (8.7) 101 48 4 (4) 5 (5) 43 (41) 5 (5) 38 19 1.12 0.20 235 74 83 20 95 33 143 12
63 13 69 (20) 26 3c 79 13 158 (46)b 46 (14) 95 45 26 (8) 9 (3) 211 (62)b 24 (7) 36 10 1.09 0.19 236 63 83 20 91 30 141 13
67 11 169 (23) 27 4b 82 14b 418 (57)b 149 (20)d 94 40 81 (11)c 27 (4) 529 (73)b 58 (8) 36 6 1.11 0.28 253 65 87 34 89 31 138 14
68 10 130 (32)b 28 4b 82 14b 238 (58)b 101 (25)b 96 43 60 (15)d 15 (4) 300 (73)b 38 (9) 36 5 1.12 0.30 270 78 91 55 85 31 137 14
67 11b 106 (29)d 28 5b 83 17b 232 (63)b 97 (27)d 95 40 61 (17)b 32 (9) 279 (75)b 58 (16)d 37 11 1.11 0.32 295 99 99 72 84 33 129 16
b
b
a Continuous data are shown as mean standard deviation and categoric data as number (%). fibrinogen 2.5 g/L.
b
p < 0.001.
b
p < 0.05.
c
p < 0.01 vs the group with
d
aPTT ¼ activated partial thromboplastin time; BMI ¼ body mass index; CABG ¼ coronary artery bypass grafting; ECC ¼ extracorporeal circulation; eGFR ¼ estimated glomerular filtration rate; INR ¼ internationalized normalized ratio; PT ¼ prothrombin time.
significant anemia occurs. Platelets were transfused in patients with ongoing bleeding exceeding 200 mL/h and a platelet count below 75 109/L, or confirmed or suspected platelet dysfunction (ie, preoperative treatment with antiplatelet drugs). Plasma was transfused in patients with ongoing bleeding exceeding 200 mL/h and signs of impaired coagulation on thromboelastometry. Prolonged clotting time in the absence of sustained heparin effect on thromboelastometry indicates a coagulation factor deficiency.
Statistics Descriptive statistics, such as mean standard deviation or median with interquartile ranges (IQR), are given for continuous variables, and proportions (%) are given for categoric variables. Categoric variables were compared between the groups with the c2 test and continuous variables with one-way analysis of variance. The importance of fibrinogen for postoperative bleeding and transfusion was analyzed (1) by handling the explanatory and the outcome variable as continuous variables, (2) by univariate testing and multiple models with fibrinogen concentration as a continuous variable and bleeding as a dichotomous variable (<1,000 mL/12 hours or >1,000 mL/12 hours), and (3) the patients were arbitrarily classified into subgroups by preoperative fibrinogen concentration. Bleeding/transfusions were compared between the different subgroups. To explore risk factors for bleeding exceeding 1,000 mL/12 hours and RBC transfusion, univariate and multiple logistic regression was performed and presented as
odds ratios (ORs) with 95% confidence intervals (CIs). All factors univariately associated with bleeding and RBC transfusion were included in the multiple models. Any p value of less than 0.05 was considered statistically significant. SPSS 20.0 software (IBM Corp, Armonk, NY) was used for all statistical analyses.
Results All Patients FIBRINOGEN CONCENTRATION. Mean standard deviation and median (IQR) for fibrinogen concentration in the entire patient cohort was 3.8 1.0 g/L and 3.6 g/L (IQR, 3.1 to 4.3 g/L), respectively. Eight patients (0.41%) had fibrinogen levels below the lower normal limit (2.0 g/L), and 370 patients (19%) had fibrinogen levels above the upper normal limit (4.5 g/L). Patients with a low fibrinogen concentration were younger, had lower weight and body mass index, lower prevalence of diabetes and smoking, lower frequency of aspirin use, and were less often operated on with CABG compared with patients with higher fibrinogen concentrations (Table 2). BLEEDING. Mean and median postoperative bleeding volume was 563 374 mL/12 hours, and 453 mL/12 hours (IQR, 340 to 650 mL/12 hours), respectively. Bleeding resulted in 96 patients (4.9%) undergoing reexplorations. There was a weak inverse correlation between postoperative blood loss and preoperative fibrinogen concentration when fibrinogen and bleeding were handled as continuous variables (r ¼ –0.17, p < 0.001).
ADULT CARDIAC
Fibrinogen Concentration, g/L a
1202
ET AL WALDEN FIBRINOGEN AND POSTOPERATIVE BLEEDING
ADULT CARDIAC
Bleeding in 205 patients (10%) exceeded 1,000 mL/12 hours. Patient characteristics and univariate ORs for patients bleeding more or less than 1,000 mL/12 hours are given in Table 3. In the multiple model, a higher preoperative fibrinogen concentration remained associated with a reduced risk of bleeding of more than 1,000 mL/12 hours (OR, 0.82; 95% CI, 0.69 to 0.97 per g/ L; p ¼ 0.024), together with higher body mass index (OR, 0.93; 95% CI, 0.88 to 0.96; p < 0.001), whereas longer extracorporeal circulation time (OR, 1.007; 95% CI, 1.004 to 1.01; p < 0.001) per minute, clopidogrel in less than 5 days before the operation (OR, 1.79; 95% CI, 1.05 to 3.07; p ¼ 0.033), low preoperative platelet count (OR, 1.002; 95% CI, 1.0 to 1.005; p ¼ 0.030), and acute operation (OR, 2.23; 95% CI, 1.17 to 4.27, p ¼ 0.015) increased the risk for bleeding more than 1,000 mL/12 hours. Other aspects of bleeding are illustrated in Figure 1, where the patients are divided into subgroups according to preoperative fibrinogen concentration. In those with lower fibrinogen levels, mean postoperative bleeding volume (Fig 1A) and the percentage of patients with bleeding exceeding 1,000 mL/12 hours (Fig 1B) were greater, whereas reoperation for bleeding was more common in patients with the lowest and highest fibrinogen levels (Fig 1C). In Table 4, the unadjusted and adjusted ORs for bleeding exceeding 1,000 mL/h are given for the different patient subgroups. After correction of all factors significantly associated with postoperative bleeding exceeding 1,000 mL/12 hours, the 2.5 g/L or less subgroup still had statistically
Ann Thorac Surg 2014;97:1199–206
significant increased risk of bleeding exceeding 1,000 mL/12 hours compared with patients with fibrinogen concentration exceeding 4.5 g/L (Table 4). TRANSFUSIONS. A total of 990 patients (51%) received RBC transfusions, 454 (23%) received plasma transfusions, and 298 (15%) received platelet transfusions. Mean and median units of RBC transfusions were 2.7 5.5 and 1 (IQR, 0 to 3), respectively, in all patients and 5.4 6.7 and 3 (IQR, 2 to 6) in transfused patients. Mean and median total numbers of blood products in all patients were 4.8 12.1 units and 1 unit (IQR, 0 to 4 units), respectively. A weak correlation was found between the preoperative fibrinogen concentration and the number of RBC transfusions during the hospital stay when the fibrinogen concentration and number of RBC transfusions were handled as continuous variables (r ¼ 0.14, p < 0.001). Patient characteristics and univariate ORs for patients receiving PRBCs are given in Table 5. In the multiple model, reduced preoperative fibrinogen concentration was not associated with an increased risk of PRBC transfusion (OR, 1.10; 95% CI, 0.89 to 1.28 per g/L; p ¼ 0.49). Acute operation (OR, 3.91; 95% CI, 1.16 to 7.98; p ¼ 0.005), reduced estimated glomerular filtration rate (OR, 1.01; 95% CI, 1.00 to 1.02 per mL/min; p ¼ 0.037), longer extracorporeal circulation time (OR, 1.02; 95% CI, 1.01 to 1.03 per min; p < 0.001), low preoperative hemoglobin concentration (OR, 1.05; 95% CI, 1.04 to 1.07 per g/L; p < 0.001), and clopidogrel in less than 5 days before the operation (OR, 2.75; 95% CI, 1.53 to 4.95; p ¼ 0.001) were associated with increased risk of PRBC transfusion in the multiple model.
Table 3. Univariate Risk Factors for Bleeding 1,000 or >1,000 mL/12 Hours (All Patients [N ¼ 1,954]) Bleeding, mL/12 Hours Variablea
1,000 (n ¼ 1,749)
Age, y Female gender BMI, kg/m2 Smoking Diabetes Pre-op medication Aspirin Clopidogrel <5 days pre-op Fibrinogen, g/L aPTT, s PT (INR) Platelet count, 109/L eGFR, mL/min Hemoglobin, g/L CABG Acute operation ECC time, min a
>1,000 (n ¼ 205)
OR (95% CI)
66 12 450 (26) 27.1 4.3 204 (12) 365 (21)
67 39 25.6 24 37
12 (19) 4.0 (12) (18)
1.00 0.67 0.91 1.02 0.84
1,220 (70) 156 (9) 3.8 1.0 36.1 9.1 1.1 0.3 262 77 88 31 137 15 980 (56) 71 (4) 94 41
142 (69) 27 (13) 3.6 1.0 36.5 5.2 1.1 0.2 245 88 84 32 136 15 95 (46) 17 (8) 110 53
0.97 1.55 0.75 1.00 1.04 0.997 0.996 0.99 0.68 2.14 1.01
p Value
(0.99–10.2) (0.47–0.98) (0.88–0.95) (0.65–1.59) (0.58–1.23)
0.33 0.037 <0.001 0.95 0.38
(0.71–1.33) (1.001–2.39) (0.64–0.89) (0.99–1.02) (0.63–1.74) (0.995–0.999) (0.99–1.001) (0.98–1.01) (0.51–0.91) (1.23–3.70) (1.004–1.010)
0.87 0.0496 0.001 0.47 0.87 0.004 0.092 0.52 0.009 0.007 <0.001
Continuous data are shown as mean standard deviation and categoric data as number (%).
aPTT ¼ activated partial thromboplastin time; interval; ECC ¼ extracorporeal circulation; odds ratio; PT ¼ prothrombin time.
BMI ¼ body mass index; CABG ¼ coronary artery bypass grafting; CI ¼ confidence eGFR ¼ estimated glomerular filtration rate; INR ¼ internationalized normalized ratio; OR ¼
1203
Fig 1. (A) Unadjusted mean postoperative blood loss, (B) blood loss exceeding 1,000 mL/12 hours, (C) reexploration rate, (D) and proportion of patients receiving transfusion of packed red blood cells in patients with different preoperative fibrinogen concentrations. All patients (N ¼ 1,954). The error bars indicate the standard deviation.
The percentage of patients receiving PRBC transfusion in the subgroups with different preoperative fibrinogen concentrations is shown in Fig 1D. In Table 6, the unadjusted and adjusted ORs for PRBC transfusion are reported for the different subgroups. Univariate testing showed patients with fibrinogen concentrations of 4.6 g/ L or more had a significantly increased risk compared with patients with concentrations of 2.5 g/L or less, but the difference was not significant in the multiple model (Table 5).
CABG Population BLEEDING. Mean and median fibrinogen concentration in the 1,075 CABG patients was 3.9 1.0 g/L and 3.7 g/L (IQR, 3.2 to 4.4 g/L), respectively. Mean and median postoperative bleeding volume was 560 351 mL/12 hours and 470 mL/12 hours (IQR, 350 to 640 mL/12 hours), respectively. Bleeding resulted in reexploration in 52 CABG patients (4.8%), and 95 CABG patients (8.8%) bled more than 1,000 mL/12 hours. In a multiple model, higher
Table 4. Unadjusted and Adjusted Odds Ratios for Bleeding >1,000 mL/12 Hours in Subgroups With Different Fibrinogen Levels Fibrinogen (g/L)
Univariate OR (95% CI)
p Value
2.5 2.6–3.0 3.1–3.8 3.9–4.5 4.6
2.92 1.97 1.68 1.20 1
0.001 0.009 0.028 0.51 –
(1.55–5.53) (1.18–3.27) (1.06–2.66) (0.70–2.03)
CI ¼ confidence interval;
Multiple Model OR (95% CI) 2.31 1.64 1.55 1.17 1
OR ¼ odds ratio.
(1.16–4.59) (0.95–2.83) (0.96–2.5) (0.68–2.03)
p Value 0.017 0.074 0.075 0.57 –
preoperative fibrinogen concentration was associated with a decreased risk of bleeding more than 1,000 mL/12 hours (OR, 0.62; 95% CI, 0.47 to 0.82 per g/L; p ¼ 0.001). CABG patients with a fibrinogen concentration of 2.5 g/L or less had a fivefold higher risk of bleeding more than 1,000 mL/12 hours than patients with a concentration of 4.6 g/L or more (OR, 5.16; 95% CI, 1.60 to 16.65; p ¼ 0.006) in univariate testing (Fig 2B). The difference remained statistically significant after correction for all factors univariately associated with bleeding more than 1,000 mL/12 hours (OR, 4.31; 95% CI, 1.27 to 14.50; p ¼ 0.019). TRANSFUSIONS. A total of 476 CABG patients received PRBC transfusions (44%), 183 (17%) received plasma transfusions, and 132 (12%) received platelet transfusions. Mean and median units of RBC transfusions were 2.0 4.1 and 0 (IQR, 0 to 2) in all CABG patients and 4.6 5.2 and 3 (IQR, 2 to 5) in CABG patients who received a transfusion. Mean and median total numbers of blood products in all CABG patients were 3.4 8.7 units and 0 units (IQR, 0 to 3 units), respectively. A high preoperative fibrinogen concentration was univariately associated with PRBC transfusion in the CABG subgroup (OR, 1.37; 95% CI, 1.21 to 1.56 per g/L; p < 0.001) but not in a multiple model (OR, 0.99; 95% CI, 0.78 to 1.25 per g/L; p ¼ 0.93).
Comment The main finding of this study was that a preoperative fibrinogen plasma concentration of less than 2.5 g/L was independently associated with increased postoperative bleeding, whereas fibrinogen concentration was not associated with RBC transfusion. The plasma concentration of fibrinogen is dependent on hereditary and acquired factors, including age, gender,
ADULT CARDIAC
ET AL WALDEN FIBRINOGEN AND POSTOPERATIVE BLEEDING
Ann Thorac Surg 2014;97:1199–206
1204
ET AL WALDEN FIBRINOGEN AND POSTOPERATIVE BLEEDING
Ann Thorac Surg 2014;97:1199–206
Table 5. Univariate Risk Factors for Blood Transfusion (All Patients [N ¼ 1,954]) Variablea ADULT CARDIAC
No Red Cell Transfusion (n ¼ 964)
Red Cell Transfusion (n ¼ 990)
64 12 158 (16) 27.4 4 106 (11) 183 (19)
68 11 331 (33) 26.5 4 124 (13) 219 (22)
682 (71) 58 (6)
680 (69) 125 (13)
3.7 0.9 36 8 1.1 0.2 257 70 96 31 143 12 599 (62) 24 (2.5) 84 31
4.0 1.1 37 10 1.1 0.3 263 85 80 30 131 15 476 (48) 64 (6.5) 106 48
Age, y Female gender BMI, kg/m2 Smoking Diabetes Pre-op medication Aspirin Clopidogrel <5 days pre-op Pre-op analyses Fibrinogen, g/L aPTT, s PT (INR) Platelet count, 109/L eGFR, mL/min Hemoglobin, g/L CABG Acute operation ECC time, min a
OR (95% CI)
p Value
(1.02–1.04) (2.06–3.18) (0.93–0.97) (0.90–1.57) (0.98–1.52)
<0.001 <0.001 <0.001 0.223 0.073
0.91 (0.75–1.11) 2.26 (1.63–3.12)
0.36 <0.001
1.03 2.57 0.95 1.19 1.22
1.36 1.02 1.64 1.001 0.98 0.94 0.56 2.71 1.015
(1.24–1.49) (1.00–1.03) (1.15–2.34) (1.000–1.002) (0.97–0.99) (0.93–0.95) (0.47–0.68) (1.68–4.36) (1.012–1.018)
<0.001 0.007 0.006 0.095 <0.001 <0.001 <0.001 <0.001 <0.001
Continuous data are shown as mean standard deviation and categoric data as number (%).
aPTT ¼ activated partial thromboplastin time; interval; ECC ¼ extracorporeal circulation; odds ratio; PT ¼ prothrombin time.
BMI ¼ body mass index; CABG ¼ coronary artery bypass grafting; CI ¼ confidence eGFR ¼ estimated glomerular filtration rate; INR ¼ internationalized normalized ratio; OR ¼
and smoking [14]. Hyperfibrinogenemia is a well-known risk factor for cardiovascular disease [14]. In addition, fibrinogen is an acute-phase reactant, and inflammatory diseases and surgical interventions increase plasma levels considerably [9, 14]. Fibrinogen levels below the lower normal level (2.0 g/L) are rare in patients without ongoing traumatic or surgical bleeding. This was confirmed in the present study, where only 0.4% of the patients had a preoperative concentration of less than 2.0 g/L. Studies in patients with hereditary hypofibrinogenemia or afibrinogenemia suggest that a fibrinogen concentration of 1.0 to 1.5 g/L is enough to avoid spontaneous bleeding episodes, and it has also been suggested that these levels are sufficient in bleeding patients without hereditary fibrinogen deficiency [15, 16].
Table 6. Unadjusted and Adjusted Odds Ratios for Receiving a Blood Transfusion in Subgroups With Different Fibrinogen Levels Fibrinogen, g/L
Univariate OR (95% CI)
p Value
2.5 2.6–3.0 3.1–3.8 3.9–4.5 4.6
0.94 1.16 1.43 2.31
1 (0.61–1.46) (0.77–1.75) (0.93–2.20) (1.49–3.59)
– 0.79 0.48 0.11 <0.001
CI ¼ confidence interval,
Multiple Model OR (95% CI)
1.20 1.24 1.31 1.46
OR ¼ odds ratio.
1 (0.57–2.55) (0.60–2.53) (0.61–2.83) (0.65–3.25)
p Value – 0.63 0.56 0.48 0.36
However, recent data suggest that when the hemostatic system is challenged by major surgical or traumatic bleeding, the fibrinogen concentration may be more important than previously assumed. Firstly, evidence from cardiac and scoliosis surgical patients indicates that there may be an inverse correlation between fibrinogen plasma concentration and the amount of perioperative bleeding [1–7, 17]. Secondly, the perioperative administration of fibrinogen concentrate reduces bleeding and the number of transfusions in diverse major surgical procedures [18–20]. Most of the previous studies have been performed in small, selected patient groups without the possibility of adjusting for other risk factors for excessive bleeding. In the present study with almost 2,000 patients, a weak but statistically significant correlation between preoperative fibrinogen plasma concentration and postoperative blood loss volume was observed. The correlation is too weak to draw any clinically meaningful conclusions. In contrast, when the association between fibrinogen concentration and the prevalence of patients bleeding more than 1,000 mL/12 hours was investigated, the picture became clearer. We found that a reduction in fibrinogen levels by 1 unit increased by approximately 25% the risk of bleeding more than 1,000 mL and that this increased risk was maintained after correction for other factors that influence postoperative bleeding. Furthermore, we found that the risk of bleeding more 1,000 mL/12 hours in patients with fibrinogen levels of less than 2.5 g/L was almost threefold higher than in patients with levels of 4.6 g/L or more (Fig 1B; Table 4). The difference was even
ET AL WALDEN FIBRINOGEN AND POSTOPERATIVE BLEEDING
1205
Fig 2. (A) Unadjusted postoperative blood loss, (B) blood loss exceeding 1,000 mL/12 hours, (C) reexploration rate, and (D) proportion of 1,075 coronary artery bypass grafting patients with different preoperative fibrinogen concentrations receiving transfusion of packed red blood cells. The error bars indicate the standard deviation.
greater in the CABG population (Fig 2B). The study therefore supports the theory that low plasma fibrinogen is associated with an increased risk of excessive postoperative bleeding. Even if fibrinogen concentration emerged as an independent predictor for bleeding more than 1,000 mL/12 hours in the present study, the relative importance of fibrinogen concentration is difficult to evaluate. The results from the multiple analyses suggest that other factors, such as acute operation, on-going dual antiplatelet therapy, and operation time, are more important than fibrinogen levels. Blood transfusions are sometimes life saving, but they also mean risks for the recipients, including acute lung injury, immunomodulation, and transfer of pathogens [21]. In the present study, univariate analysis indicated that high fibrinogen levels increase the risk of RBC transfusion. However, it is apparent from Table 2 that patients with higher fibrinogen levels were older, were more likely to be female, had lower preoperative hemoglobin levels, and were more likely to be operated on acutely. All of these factors are well-known risk factors for transfusion requirements. Accordingly, after adjustment for confounding factors, preoperative fibrinogen concentration was not associated with an increased risk of RBC transfusion. Thus, our results suggest that the preoperative fibrinogen concentration has no or minimal influence on transfusion requirements in cardiac operations. This contrasts with a study by Prohaska and colleagues [22] involving 2,831 cardiac surgical patients where a high preoperative fibrinogen concentration was independently associated with PRBC transfusion. However, the studies differ in several important aspects (eg, prospective/retrospective, different factors included in the multiple model). Two smaller studies have found
an association between low plasma fibrinogen and increased blood usage after cardiac operations [1, 2]. Interestingly, postoperative bleeding volume did not emerge as predictor for PRBC transfusion in the multiple model. This underlines that the decision to administer transfusions to cardiac surgical patients is based on multiple factors besides bleeding volume. However, several of the variables predicting transfusion were also related to postoperative bleeding volume and low fibrinogen levels, which may obscure the analysis. In the absence of a generally accepted definition of excessive bleeding in cardiac surgery, we defined arbitrarily excessive bleeding as more than 1,000 mL/12 hours, which is twice the mean volume in the present study and involved 10% of the patients. Previous studies have used different definitions of excessive bleeding. Another potential limitation was that patients who underwent operations on the ascending aorta were excluded. In conclusion, this study found a preoperative plasma fibrinogen concentration of less than 2.5 g/L was associated with increased risk of excessive bleeding after cardiac operations. Fibrinogen concentration did not influence RBC transfusion requirements. This work was supported by V€astra G€ otaland region, the Gothenburg Medical Society, and the Swedish Heart and Lung Foundation. The sponsors of the study had no influence on the analysis and interpretation of data, in the writing of the report, or in the decision to submit the paper for publication.
References 1. Josefy S, Briones R, Bryant BJ. Preoperative coagulation studies to predict blood component usage in coronary artery bypass graft surgery. Immunohematology 2011;27:151–3.
ADULT CARDIAC
Ann Thorac Surg 2014;97:1199–206
1206
ET AL WALDEN FIBRINOGEN AND POSTOPERATIVE BLEEDING
ADULT CARDIAC
2. Karlsson M, Ternstrom L, Hyllner M, Baghaei F, Nilsson S, Jeppsson A. Plasma fibrinogen level, bleeding, and transfusion after on-pump coronary artery bypass grafting surgery: a prospective observational study. Transfusion 2008;48:2152–8. 3. Aljassim O, Karlsson M, Wiklund L, Jeppsson A, Olsson P, Berglin E. Inflammatory response and platelet activation after off-pump coronary artery bypass surgery. Scand Cardiovasc J 2006;40:43–8. 4. Blome M, Isgro F, Kiessling AH, et al. Relationship between factor XIII activity, fibrinogen, haemostasis screening tests and postoperative bleeding in cardiopulmonary bypass surgery. Thromb Haemost 2005;93:1101–7. 5. Gravlee GP, Arora S, Lavender SW, et al. Predictive value of blood clotting tests in cardiac surgical patients. Ann Thorac Surg 1994;58:216–21. 6. Wahba A, Rothe G, Lodes H, Barlage S, Schmitz G, Birnbaum DE. Predictors of blood loss after coronary artery bypass grafting. J Cardiothorac Vasc Anesth 1997;11:824–7. 7. Liu G, McNicol PL, McCall PR, et al. Prediction of the mediastinal drainage after coronary artery bypass surgery. Anaesth Intensive Care 2000;28:420–6. 8. Bolliger D, Gonsahn M, Levy JH, Williams WH, Tanaka KA. Is preoperative fibrinogen predictive for postoperative bleeding after coronary artery bypass grafting surgery? Transfusion 2009;49:2006–7. 9. Ternstrom L, Radulovic V, Karlsson M, et al. Plasma activity of individual coagulation factors, hemodilution and blood loss after cardiac surgery: a prospective observational study. Thromb Res 2010;126:e128–33. 10. Emeklibas N, Kammerer I, Bach J, Sack FU, Hellstern P. Preoperative hemostasis and its association with bleeding and blood component transfusion requirements in cardiopulmonary bypass surgery. Transfusion 2013;53:1226–34. 11. Poston R, Gu J, Manchio J, et al. Platelet function tests predict bleeding and thrombotic events after off-pump coronary bypass grafting. Eur J Cardiothorac Surg 2005;27:584–91. 12. Clauss A. [Rapid physiological coagulation method in determination of fibrinogen]. Acta Haematol 1957;17:237–46.
Ann Thorac Surg 2014;97:1199–206
13. Cockcroft DW, Gault MH. Prediction of creatinine clearance from serum creatinine. Nephron 1976;16:31–41. 14. The Fibrinogen Studies Collaboration. Associations of plasma fibrinogen levels with established cardiovascular disease risk factors, inflammatory markers, and other characteristics: Individual participant meta-analysis of 154,211 adults in 31 prospective studies. Am J Epidemiol 2007;166: 867–79. 15. Hellstern P, Muntean W, Schramm W, Seifried E, Solheim BG. Practical guidelines for the clinical use of plasma. Thromb Res 2002;107:S53–7. 16. Stainsby D, MacLennan S, Thomas D, Isaac J, Hamilton PJ. Guidelines on the management of massive blood loss. Br J Haematol 2006;135:634–41. 17. Carling MS, Jeppsson A, Wessberg P, Henriksson A, Baghaei F, Brisby H. Preoperative fibrinogen plasma concentration is associated with perioperative bleeding and transfusion requirements in scoliosis surgery. Spine 2011;36: 549–55. 18. Karlsson M, Ternstrom L, Hyllner M, et al. Prophylactic fibrinogen infusion reduces bleeding after coronary artery bypass surgery. A prospective randomised pilot study. Thromb Haemost 2009;102:137–44. 19. Fenger-Eriksen C, Lindberg-Larsen M, Christensen AQ, Ingerslev J, Sorensen B. Fibrinogen concentrate substitution therapy in patients with massive haemorrhage and low plasma fibrinogen concentrations. Br J Anaesth 2008;101: 769–73. 20. Rahe-Meyer N, Solomon C, Hanke A, et al. Effects of fibrinogen concentrate as first-line therapy during major aortic replacement surgery: a randomized, placebocontrolled trial. Anesthesiology 2013;118:40–50. 21. Buddeberg F, Schimmer BB, Spahn DR. Transfusion-transmissible infections and transfusion-related immunomodulation. Best Pract Res Clin Anaesthesiol 2008;22:503–17. 22. Prohaska W, Zittermann A, Inoue K, et al. Preoperative haemostasis testing does not predict requirement of blood products in cardiac surgery. Eur J Med Res 2008;13:525–30.
Background. Data from small selected patient populations suggest that the preoperative plasma concentration of fibrinogen influences postoperative blood loss and red blood cell transfusion after cardiac operations, but there are also conflicting reports. We assessed the importance of preoperative fibrinogen concentration for excessive bleeding and red cell blood transfusion in a large cohort of mixed cardiac surgical patients. Methods. We included 1,954 cardiac surgical patients in a prospective observational study. The fibrinogen plasma concentration was measured on the day before the operation. Blood loss (mediastinal drain volume) during the first 12 postoperative hours and red blood cell transfusion during the hospital stay were registered and related to fibrinogen concentration with logistic regression models. Excessive bleeding was defined as postoperative blood loss exceeding 1,000 mL/12 hours. Results. The preoperative fibrinogen concentration was inversely proportional to the prevalence of excessive
bleeding in univariate testing (odds ratio [OR], 0.75; 95% confidence interval [CI], 0.64 to 0.89 per g/L; p [ 0.001) and also in a multiple model adjusted for age, sex, body mass index, renal function, acuteness of the operation, cardiopulmonary bypass time, clopidogrel use less than 5 days before the operation, and type of operation (OR for fibrinogen, 0.82; 95% CI, 0.69 to 0.97; p [ 0.024). In contrast, the prevalence of red cell blood transfusion increased with increasing fibrinogen levels in univariate testing (OR, 1.36; 95% CI, 1.24 to 1.49; p < 0.001) but not in a multiple model (OR, 1.10; 95% CI, 0.89 to 1.28; p [ 0.49). Conclusions. Preoperative plasma concentration of fibrinogen is independently associated with excessive bleeding after cardiac operations but not with red blood cell transfusion.
F
Patients and Methods
ibrinogen is a key factor in the coagulation cascade. There are conflicting data on whether the preoperative plasma concentration of fibrinogen is associated with postoperative bleeding volume and transfusion after cardiac operations. Although some studies have found a statistically significant correlation with bleeding [1–7], contradictory reports exist [8–11]. Most of the previous studies have been too small to allow adjustment for confounding factors. In addition, most of them have included only elective coronary artery bypass grafting (CABG) patients. We sought to determine the relationship between the preoperative plasma fibrinogen concentration, on the one hand, and postoperative bleeding and red blood cell (RBC) transfusions, on the other, in a large unselected cohort of mixed cardiac surgical patients. Accepted for publication Nov 25, 2013. Address correspondence to Dr Jeppsson, Department of Cardiothoracic Surgery, Sahlgrenska University Hospital, 413 45 Gothenburg, Sweden; e-mail: [email protected].
Ó 2014 by The Society of Thoracic Surgeons Published by Elsevier Inc
(Ann Thorac Surg 2014;97:1199–206) Ó 2014 by The Society of Thoracic Surgeons
This study was approved by the Regional Research Ethics Committee, which waived the requirement for individual patient consent.
Patients From February 2009 to January 2011, 2,208 adult cardiac surgical procedures were performed at our institution. We excluded 170 (6.1%) procedures due to missing data, resulting in a study cohort of 1,954 procedures. All types of acute and elective procedures were included: CABG, valve operations, combined CABG and valve operations, adult congenital operations, and arrhythmia operations. Patients who underwent operations on the ascending aorta, including dissections and aneurysms, were excluded because of their high risk of surgical bleeding. If a patient underwent more than one operation during a single hospital admission, only the first operation was included in the analysis. A separate analysis was performed in isolated CABG patients to allow comparisons 0003-4975/$36.00 http://dx.doi.org/10.1016/j.athoracsur.2013.11.064
ADULT CARDIAC
CARDIOTHORACIC ANESTHESIOLOGY: The Annals of Thoracic Surgery CME Program is located online at http://www.annalsthoracicsurgery.org/cme/ home. To take the CME activity related to this article, you must have either an STS member or an individual non-member subscription to the journal.
1200
ET AL WALDEN FIBRINOGEN AND POSTOPERATIVE BLEEDING
Ann Thorac Surg 2014;97:1199–206
ADULT CARDIAC
with previous studies. Patient characteristics are given in Table 1. To further illustrate the importance of fibrinogen concentration, the patients were arbitrarily divided into five groups according to the preoperative fibrinogen concentration (2.5, 2.6 to 3.0, 3.1 to 3.8, 3.9 to 4.5, and 4.6 g/L). Patient characteristics for each group are reported in Table 2.
Laboratory Variables Blood samples were collected on the day before the operation (elective cases) or immediately before the operation (acute cases). Plasma fibrinogen concentration was determined according to the method of Clauss [12] (STA-FIB 2; Diagnostica Stago, Asnieres, France). The reference value is 2.0 to 4.5 g/L. Activated partial thromboplastin time, prothrombin time, serum creatinine, and hemoglobin were measured preoperatively with standard clinical methods. Estimated glomerular filtration rate was calculated with the Cockcroft-Gault formula [13].
Clinical Management Anesthesia in all patients was induced with fentanyl, thiopentone, and pancuronium and maintained with sevoflurane or propofol. The patients received heparin (300 U/kg body weight) to maintain an activated clotting
time exceeding 480 seconds. After cardiopulmonary bypass, the heparin was reversed by the administration of 1 mg protamine/100 U heparin. A cell-saving device was not used routinely. Cardioprotection was achieved with antegrade cold blood cardioplegia. Thromboelastometry/ graphy was not used routinely. In patients with on-going bleeding, thromboelastometry was used to assess coagulation and guide transfusions of blood products. Venous reservoir blood was not processed in a cell-saving device before retransfusion. Aspirin was not discontinued preoperatively. In elective cases, clopidogrel was discontinued 5 days before surgery. All patients received 2 g tranexamic acid (Cyklokapron; Pfizer Inc, New York, NY) preoperatively and 2 g postoperatively after weaning off cardiopulmonary bypass. Aprotinin was not used.
Bleeding and Transfusions Postoperative bleeding was defined as the total amount of chest tube drainage after closure of the sternum and during the first 12 postoperative hours, or until the patient was reexplored because of bleeding. The total amount of transfused packed RBCs (PRBCs), fresh frozen plasma, and platelets during the hospital stay was recorded. The local protocol specifies that PRBC transfusions should be given when the blood hemoglobin level has decreased to below 70 g/L or when a clinically
Table 1. Characteristics for All Patients and for the Subgroups With and Without Coronary Artery Bypass Grafting Variablea Age, y Female gender BMI, kg/m2 Smoking Diabetes Preoperative medication Aspirin Clopidogrel (<5 days pre-op) Fibrinogen, g/L aPTT, s PT (INR) Platelet count, 109/L Plasma creatinine, mmol/L eGFR, mL/min Hemoglobin, g/L Operative procedure CABG Aortic valve Mitral valve CABG þ valve Other Acute operation ECC time, min a
All Patients (N ¼ 1,954)
CABG Patients (n ¼ 1,075)
Non-CABG Patients (n ¼ 879)
p Value (CABG vs non-CABG)
66 12 489 (25) 27 4 228 (12) 402 (21)
67 9 235 (22) 27 4 154 (14) 278 (26)
64 14 254 (29) 26 4 74 (8) 124 (14)
<0.001 <0.001 <0.001 <0.001 <0.001
1,362 (70) 183 (9) 3.8 1.0 36 9 1.1 0.3 260 78 89 46 88 31 137 15
1,054 (98) 176 (16) 3.9 1.0 36 10 1.1 0.2 267 75 89 40 88 30 138 14
308 (35) 7 (1) 3.7 1.0 36 7 1.2 0.3 252 81 90 54 88 34 136 16
<0.001 <0.001 <0.001 0.69 <0.001 <0.001 0.54 0.82 0.004
1,075 (55) 387 (20) 154 (8) 205 (10) 133 (7) 88 (4.5) 95 42
1,075 – – – – 66 (6.1) 77 27
– 387 (44) 154 (18) 205 (23) 133 (15) 22 (2.5) 117 46
– – – – – <0.001 <0.001
Continuous data are shown as mean standard deviation and categoric data as number (%).
aPTT ¼ activated partial thromboplastin time; BMI ¼ body mass index; CABG ¼ coronary artery bypass grafting; ECC ¼ extracorporeal circulation; eGFR ¼ estimated glomerular filtration rate; INR ¼ internationalized normalized ratio; PT ¼ prothrombin time.
ET AL WALDEN FIBRINOGEN AND POSTOPERATIVE BLEEDING
Ann Thorac Surg 2014;97:1199–206
1201
Table 2. Characteristics Related to Preoperative Fibrinogen Concentration in All Patients (N ¼ 1,954)
Variable
Age, y Female gender BMI, kg/m2 Weight, kg CABG Diabetes ECC time, min Smoking Acute operation Aspirin Clopidogrel (<5 days pre-op) APTT, s PT (INR) Platelet count, 109/L Serum creatinine, mmol/L eGFR, mL/min Hemoglobin, g/L
2.5 (n ¼ 105)
2.6–3.0 (n ¼ 340)
3.1–3.8 (n ¼ 729)
3.9–4.5 (n ¼ 410)
4.6 (n ¼ 370)
57 18 16 (15) 25 4 76 13 29 (28) 9 (8.7) 101 48 4 (4) 5 (5) 43 (41) 5 (5) 38 19 1.12 0.20 235 74 83 20 95 33 143 12
63 13 69 (20) 26 3c 79 13 158 (46)b 46 (14) 95 45 26 (8) 9 (3) 211 (62)b 24 (7) 36 10 1.09 0.19 236 63 83 20 91 30 141 13
67 11 169 (23) 27 4b 82 14b 418 (57)b 149 (20)d 94 40 81 (11)c 27 (4) 529 (73)b 58 (8) 36 6 1.11 0.28 253 65 87 34 89 31 138 14
68 10 130 (32)b 28 4b 82 14b 238 (58)b 101 (25)b 96 43 60 (15)d 15 (4) 300 (73)b 38 (9) 36 5 1.12 0.30 270 78 91 55 85 31 137 14
67 11b 106 (29)d 28 5b 83 17b 232 (63)b 97 (27)d 95 40 61 (17)b 32 (9) 279 (75)b 58 (16)d 37 11 1.11 0.32 295 99 99 72 84 33 129 16
b
b
a Continuous data are shown as mean standard deviation and categoric data as number (%). fibrinogen 2.5 g/L.
b
p < 0.001.
b
p < 0.05.
c
p < 0.01 vs the group with
d
aPTT ¼ activated partial thromboplastin time; BMI ¼ body mass index; CABG ¼ coronary artery bypass grafting; ECC ¼ extracorporeal circulation; eGFR ¼ estimated glomerular filtration rate; INR ¼ internationalized normalized ratio; PT ¼ prothrombin time.
significant anemia occurs. Platelets were transfused in patients with ongoing bleeding exceeding 200 mL/h and a platelet count below 75 109/L, or confirmed or suspected platelet dysfunction (ie, preoperative treatment with antiplatelet drugs). Plasma was transfused in patients with ongoing bleeding exceeding 200 mL/h and signs of impaired coagulation on thromboelastometry. Prolonged clotting time in the absence of sustained heparin effect on thromboelastometry indicates a coagulation factor deficiency.
Statistics Descriptive statistics, such as mean standard deviation or median with interquartile ranges (IQR), are given for continuous variables, and proportions (%) are given for categoric variables. Categoric variables were compared between the groups with the c2 test and continuous variables with one-way analysis of variance. The importance of fibrinogen for postoperative bleeding and transfusion was analyzed (1) by handling the explanatory and the outcome variable as continuous variables, (2) by univariate testing and multiple models with fibrinogen concentration as a continuous variable and bleeding as a dichotomous variable (<1,000 mL/12 hours or >1,000 mL/12 hours), and (3) the patients were arbitrarily classified into subgroups by preoperative fibrinogen concentration. Bleeding/transfusions were compared between the different subgroups. To explore risk factors for bleeding exceeding 1,000 mL/12 hours and RBC transfusion, univariate and multiple logistic regression was performed and presented as
odds ratios (ORs) with 95% confidence intervals (CIs). All factors univariately associated with bleeding and RBC transfusion were included in the multiple models. Any p value of less than 0.05 was considered statistically significant. SPSS 20.0 software (IBM Corp, Armonk, NY) was used for all statistical analyses.
Results All Patients FIBRINOGEN CONCENTRATION. Mean standard deviation and median (IQR) for fibrinogen concentration in the entire patient cohort was 3.8 1.0 g/L and 3.6 g/L (IQR, 3.1 to 4.3 g/L), respectively. Eight patients (0.41%) had fibrinogen levels below the lower normal limit (2.0 g/L), and 370 patients (19%) had fibrinogen levels above the upper normal limit (4.5 g/L). Patients with a low fibrinogen concentration were younger, had lower weight and body mass index, lower prevalence of diabetes and smoking, lower frequency of aspirin use, and were less often operated on with CABG compared with patients with higher fibrinogen concentrations (Table 2). BLEEDING. Mean and median postoperative bleeding volume was 563 374 mL/12 hours, and 453 mL/12 hours (IQR, 340 to 650 mL/12 hours), respectively. Bleeding resulted in 96 patients (4.9%) undergoing reexplorations. There was a weak inverse correlation between postoperative blood loss and preoperative fibrinogen concentration when fibrinogen and bleeding were handled as continuous variables (r ¼ –0.17, p < 0.001).
ADULT CARDIAC
Fibrinogen Concentration, g/L a
1202
ET AL WALDEN FIBRINOGEN AND POSTOPERATIVE BLEEDING
ADULT CARDIAC
Bleeding in 205 patients (10%) exceeded 1,000 mL/12 hours. Patient characteristics and univariate ORs for patients bleeding more or less than 1,000 mL/12 hours are given in Table 3. In the multiple model, a higher preoperative fibrinogen concentration remained associated with a reduced risk of bleeding of more than 1,000 mL/12 hours (OR, 0.82; 95% CI, 0.69 to 0.97 per g/ L; p ¼ 0.024), together with higher body mass index (OR, 0.93; 95% CI, 0.88 to 0.96; p < 0.001), whereas longer extracorporeal circulation time (OR, 1.007; 95% CI, 1.004 to 1.01; p < 0.001) per minute, clopidogrel in less than 5 days before the operation (OR, 1.79; 95% CI, 1.05 to 3.07; p ¼ 0.033), low preoperative platelet count (OR, 1.002; 95% CI, 1.0 to 1.005; p ¼ 0.030), and acute operation (OR, 2.23; 95% CI, 1.17 to 4.27, p ¼ 0.015) increased the risk for bleeding more than 1,000 mL/12 hours. Other aspects of bleeding are illustrated in Figure 1, where the patients are divided into subgroups according to preoperative fibrinogen concentration. In those with lower fibrinogen levels, mean postoperative bleeding volume (Fig 1A) and the percentage of patients with bleeding exceeding 1,000 mL/12 hours (Fig 1B) were greater, whereas reoperation for bleeding was more common in patients with the lowest and highest fibrinogen levels (Fig 1C). In Table 4, the unadjusted and adjusted ORs for bleeding exceeding 1,000 mL/h are given for the different patient subgroups. After correction of all factors significantly associated with postoperative bleeding exceeding 1,000 mL/12 hours, the 2.5 g/L or less subgroup still had statistically
Ann Thorac Surg 2014;97:1199–206
significant increased risk of bleeding exceeding 1,000 mL/12 hours compared with patients with fibrinogen concentration exceeding 4.5 g/L (Table 4). TRANSFUSIONS. A total of 990 patients (51%) received RBC transfusions, 454 (23%) received plasma transfusions, and 298 (15%) received platelet transfusions. Mean and median units of RBC transfusions were 2.7 5.5 and 1 (IQR, 0 to 3), respectively, in all patients and 5.4 6.7 and 3 (IQR, 2 to 6) in transfused patients. Mean and median total numbers of blood products in all patients were 4.8 12.1 units and 1 unit (IQR, 0 to 4 units), respectively. A weak correlation was found between the preoperative fibrinogen concentration and the number of RBC transfusions during the hospital stay when the fibrinogen concentration and number of RBC transfusions were handled as continuous variables (r ¼ 0.14, p < 0.001). Patient characteristics and univariate ORs for patients receiving PRBCs are given in Table 5. In the multiple model, reduced preoperative fibrinogen concentration was not associated with an increased risk of PRBC transfusion (OR, 1.10; 95% CI, 0.89 to 1.28 per g/L; p ¼ 0.49). Acute operation (OR, 3.91; 95% CI, 1.16 to 7.98; p ¼ 0.005), reduced estimated glomerular filtration rate (OR, 1.01; 95% CI, 1.00 to 1.02 per mL/min; p ¼ 0.037), longer extracorporeal circulation time (OR, 1.02; 95% CI, 1.01 to 1.03 per min; p < 0.001), low preoperative hemoglobin concentration (OR, 1.05; 95% CI, 1.04 to 1.07 per g/L; p < 0.001), and clopidogrel in less than 5 days before the operation (OR, 2.75; 95% CI, 1.53 to 4.95; p ¼ 0.001) were associated with increased risk of PRBC transfusion in the multiple model.
Table 3. Univariate Risk Factors for Bleeding 1,000 or >1,000 mL/12 Hours (All Patients [N ¼ 1,954]) Bleeding, mL/12 Hours Variablea
1,000 (n ¼ 1,749)
Age, y Female gender BMI, kg/m2 Smoking Diabetes Pre-op medication Aspirin Clopidogrel <5 days pre-op Fibrinogen, g/L aPTT, s PT (INR) Platelet count, 109/L eGFR, mL/min Hemoglobin, g/L CABG Acute operation ECC time, min a
>1,000 (n ¼ 205)
OR (95% CI)
66 12 450 (26) 27.1 4.3 204 (12) 365 (21)
67 39 25.6 24 37
12 (19) 4.0 (12) (18)
1.00 0.67 0.91 1.02 0.84
1,220 (70) 156 (9) 3.8 1.0 36.1 9.1 1.1 0.3 262 77 88 31 137 15 980 (56) 71 (4) 94 41
142 (69) 27 (13) 3.6 1.0 36.5 5.2 1.1 0.2 245 88 84 32 136 15 95 (46) 17 (8) 110 53
0.97 1.55 0.75 1.00 1.04 0.997 0.996 0.99 0.68 2.14 1.01
p Value
(0.99–10.2) (0.47–0.98) (0.88–0.95) (0.65–1.59) (0.58–1.23)
0.33 0.037 <0.001 0.95 0.38
(0.71–1.33) (1.001–2.39) (0.64–0.89) (0.99–1.02) (0.63–1.74) (0.995–0.999) (0.99–1.001) (0.98–1.01) (0.51–0.91) (1.23–3.70) (1.004–1.010)
0.87 0.0496 0.001 0.47 0.87 0.004 0.092 0.52 0.009 0.007 <0.001
Continuous data are shown as mean standard deviation and categoric data as number (%).
aPTT ¼ activated partial thromboplastin time; interval; ECC ¼ extracorporeal circulation; odds ratio; PT ¼ prothrombin time.
BMI ¼ body mass index; CABG ¼ coronary artery bypass grafting; CI ¼ confidence eGFR ¼ estimated glomerular filtration rate; INR ¼ internationalized normalized ratio; OR ¼
1203
Fig 1. (A) Unadjusted mean postoperative blood loss, (B) blood loss exceeding 1,000 mL/12 hours, (C) reexploration rate, (D) and proportion of patients receiving transfusion of packed red blood cells in patients with different preoperative fibrinogen concentrations. All patients (N ¼ 1,954). The error bars indicate the standard deviation.
The percentage of patients receiving PRBC transfusion in the subgroups with different preoperative fibrinogen concentrations is shown in Fig 1D. In Table 6, the unadjusted and adjusted ORs for PRBC transfusion are reported for the different subgroups. Univariate testing showed patients with fibrinogen concentrations of 4.6 g/ L or more had a significantly increased risk compared with patients with concentrations of 2.5 g/L or less, but the difference was not significant in the multiple model (Table 5).
CABG Population BLEEDING. Mean and median fibrinogen concentration in the 1,075 CABG patients was 3.9 1.0 g/L and 3.7 g/L (IQR, 3.2 to 4.4 g/L), respectively. Mean and median postoperative bleeding volume was 560 351 mL/12 hours and 470 mL/12 hours (IQR, 350 to 640 mL/12 hours), respectively. Bleeding resulted in reexploration in 52 CABG patients (4.8%), and 95 CABG patients (8.8%) bled more than 1,000 mL/12 hours. In a multiple model, higher
Table 4. Unadjusted and Adjusted Odds Ratios for Bleeding >1,000 mL/12 Hours in Subgroups With Different Fibrinogen Levels Fibrinogen (g/L)
Univariate OR (95% CI)
p Value
2.5 2.6–3.0 3.1–3.8 3.9–4.5 4.6
2.92 1.97 1.68 1.20 1
0.001 0.009 0.028 0.51 –
(1.55–5.53) (1.18–3.27) (1.06–2.66) (0.70–2.03)
CI ¼ confidence interval;
Multiple Model OR (95% CI) 2.31 1.64 1.55 1.17 1
OR ¼ odds ratio.
(1.16–4.59) (0.95–2.83) (0.96–2.5) (0.68–2.03)
p Value 0.017 0.074 0.075 0.57 –
preoperative fibrinogen concentration was associated with a decreased risk of bleeding more than 1,000 mL/12 hours (OR, 0.62; 95% CI, 0.47 to 0.82 per g/L; p ¼ 0.001). CABG patients with a fibrinogen concentration of 2.5 g/L or less had a fivefold higher risk of bleeding more than 1,000 mL/12 hours than patients with a concentration of 4.6 g/L or more (OR, 5.16; 95% CI, 1.60 to 16.65; p ¼ 0.006) in univariate testing (Fig 2B). The difference remained statistically significant after correction for all factors univariately associated with bleeding more than 1,000 mL/12 hours (OR, 4.31; 95% CI, 1.27 to 14.50; p ¼ 0.019). TRANSFUSIONS. A total of 476 CABG patients received PRBC transfusions (44%), 183 (17%) received plasma transfusions, and 132 (12%) received platelet transfusions. Mean and median units of RBC transfusions were 2.0 4.1 and 0 (IQR, 0 to 2) in all CABG patients and 4.6 5.2 and 3 (IQR, 2 to 5) in CABG patients who received a transfusion. Mean and median total numbers of blood products in all CABG patients were 3.4 8.7 units and 0 units (IQR, 0 to 3 units), respectively. A high preoperative fibrinogen concentration was univariately associated with PRBC transfusion in the CABG subgroup (OR, 1.37; 95% CI, 1.21 to 1.56 per g/L; p < 0.001) but not in a multiple model (OR, 0.99; 95% CI, 0.78 to 1.25 per g/L; p ¼ 0.93).
Comment The main finding of this study was that a preoperative fibrinogen plasma concentration of less than 2.5 g/L was independently associated with increased postoperative bleeding, whereas fibrinogen concentration was not associated with RBC transfusion. The plasma concentration of fibrinogen is dependent on hereditary and acquired factors, including age, gender,
ADULT CARDIAC
ET AL WALDEN FIBRINOGEN AND POSTOPERATIVE BLEEDING
Ann Thorac Surg 2014;97:1199–206
1204
ET AL WALDEN FIBRINOGEN AND POSTOPERATIVE BLEEDING
Ann Thorac Surg 2014;97:1199–206
Table 5. Univariate Risk Factors for Blood Transfusion (All Patients [N ¼ 1,954]) Variablea ADULT CARDIAC
No Red Cell Transfusion (n ¼ 964)
Red Cell Transfusion (n ¼ 990)
64 12 158 (16) 27.4 4 106 (11) 183 (19)
68 11 331 (33) 26.5 4 124 (13) 219 (22)
682 (71) 58 (6)
680 (69) 125 (13)
3.7 0.9 36 8 1.1 0.2 257 70 96 31 143 12 599 (62) 24 (2.5) 84 31
4.0 1.1 37 10 1.1 0.3 263 85 80 30 131 15 476 (48) 64 (6.5) 106 48
Age, y Female gender BMI, kg/m2 Smoking Diabetes Pre-op medication Aspirin Clopidogrel <5 days pre-op Pre-op analyses Fibrinogen, g/L aPTT, s PT (INR) Platelet count, 109/L eGFR, mL/min Hemoglobin, g/L CABG Acute operation ECC time, min a
OR (95% CI)
p Value
(1.02–1.04) (2.06–3.18) (0.93–0.97) (0.90–1.57) (0.98–1.52)
<0.001 <0.001 <0.001 0.223 0.073
0.91 (0.75–1.11) 2.26 (1.63–3.12)
0.36 <0.001
1.03 2.57 0.95 1.19 1.22
1.36 1.02 1.64 1.001 0.98 0.94 0.56 2.71 1.015
(1.24–1.49) (1.00–1.03) (1.15–2.34) (1.000–1.002) (0.97–0.99) (0.93–0.95) (0.47–0.68) (1.68–4.36) (1.012–1.018)
<0.001 0.007 0.006 0.095 <0.001 <0.001 <0.001 <0.001 <0.001
Continuous data are shown as mean standard deviation and categoric data as number (%).
aPTT ¼ activated partial thromboplastin time; interval; ECC ¼ extracorporeal circulation; odds ratio; PT ¼ prothrombin time.
BMI ¼ body mass index; CABG ¼ coronary artery bypass grafting; CI ¼ confidence eGFR ¼ estimated glomerular filtration rate; INR ¼ internationalized normalized ratio; OR ¼
and smoking [14]. Hyperfibrinogenemia is a well-known risk factor for cardiovascular disease [14]. In addition, fibrinogen is an acute-phase reactant, and inflammatory diseases and surgical interventions increase plasma levels considerably [9, 14]. Fibrinogen levels below the lower normal level (2.0 g/L) are rare in patients without ongoing traumatic or surgical bleeding. This was confirmed in the present study, where only 0.4% of the patients had a preoperative concentration of less than 2.0 g/L. Studies in patients with hereditary hypofibrinogenemia or afibrinogenemia suggest that a fibrinogen concentration of 1.0 to 1.5 g/L is enough to avoid spontaneous bleeding episodes, and it has also been suggested that these levels are sufficient in bleeding patients without hereditary fibrinogen deficiency [15, 16].
Table 6. Unadjusted and Adjusted Odds Ratios for Receiving a Blood Transfusion in Subgroups With Different Fibrinogen Levels Fibrinogen, g/L
Univariate OR (95% CI)
p Value
2.5 2.6–3.0 3.1–3.8 3.9–4.5 4.6
0.94 1.16 1.43 2.31
1 (0.61–1.46) (0.77–1.75) (0.93–2.20) (1.49–3.59)
– 0.79 0.48 0.11 <0.001
CI ¼ confidence interval,
Multiple Model OR (95% CI)
1.20 1.24 1.31 1.46
OR ¼ odds ratio.
1 (0.57–2.55) (0.60–2.53) (0.61–2.83) (0.65–3.25)
p Value – 0.63 0.56 0.48 0.36
However, recent data suggest that when the hemostatic system is challenged by major surgical or traumatic bleeding, the fibrinogen concentration may be more important than previously assumed. Firstly, evidence from cardiac and scoliosis surgical patients indicates that there may be an inverse correlation between fibrinogen plasma concentration and the amount of perioperative bleeding [1–7, 17]. Secondly, the perioperative administration of fibrinogen concentrate reduces bleeding and the number of transfusions in diverse major surgical procedures [18–20]. Most of the previous studies have been performed in small, selected patient groups without the possibility of adjusting for other risk factors for excessive bleeding. In the present study with almost 2,000 patients, a weak but statistically significant correlation between preoperative fibrinogen plasma concentration and postoperative blood loss volume was observed. The correlation is too weak to draw any clinically meaningful conclusions. In contrast, when the association between fibrinogen concentration and the prevalence of patients bleeding more than 1,000 mL/12 hours was investigated, the picture became clearer. We found that a reduction in fibrinogen levels by 1 unit increased by approximately 25% the risk of bleeding more than 1,000 mL and that this increased risk was maintained after correction for other factors that influence postoperative bleeding. Furthermore, we found that the risk of bleeding more 1,000 mL/12 hours in patients with fibrinogen levels of less than 2.5 g/L was almost threefold higher than in patients with levels of 4.6 g/L or more (Fig 1B; Table 4). The difference was even
ET AL WALDEN FIBRINOGEN AND POSTOPERATIVE BLEEDING
1205
Fig 2. (A) Unadjusted postoperative blood loss, (B) blood loss exceeding 1,000 mL/12 hours, (C) reexploration rate, and (D) proportion of 1,075 coronary artery bypass grafting patients with different preoperative fibrinogen concentrations receiving transfusion of packed red blood cells. The error bars indicate the standard deviation.
greater in the CABG population (Fig 2B). The study therefore supports the theory that low plasma fibrinogen is associated with an increased risk of excessive postoperative bleeding. Even if fibrinogen concentration emerged as an independent predictor for bleeding more than 1,000 mL/12 hours in the present study, the relative importance of fibrinogen concentration is difficult to evaluate. The results from the multiple analyses suggest that other factors, such as acute operation, on-going dual antiplatelet therapy, and operation time, are more important than fibrinogen levels. Blood transfusions are sometimes life saving, but they also mean risks for the recipients, including acute lung injury, immunomodulation, and transfer of pathogens [21]. In the present study, univariate analysis indicated that high fibrinogen levels increase the risk of RBC transfusion. However, it is apparent from Table 2 that patients with higher fibrinogen levels were older, were more likely to be female, had lower preoperative hemoglobin levels, and were more likely to be operated on acutely. All of these factors are well-known risk factors for transfusion requirements. Accordingly, after adjustment for confounding factors, preoperative fibrinogen concentration was not associated with an increased risk of RBC transfusion. Thus, our results suggest that the preoperative fibrinogen concentration has no or minimal influence on transfusion requirements in cardiac operations. This contrasts with a study by Prohaska and colleagues [22] involving 2,831 cardiac surgical patients where a high preoperative fibrinogen concentration was independently associated with PRBC transfusion. However, the studies differ in several important aspects (eg, prospective/retrospective, different factors included in the multiple model). Two smaller studies have found
an association between low plasma fibrinogen and increased blood usage after cardiac operations [1, 2]. Interestingly, postoperative bleeding volume did not emerge as predictor for PRBC transfusion in the multiple model. This underlines that the decision to administer transfusions to cardiac surgical patients is based on multiple factors besides bleeding volume. However, several of the variables predicting transfusion were also related to postoperative bleeding volume and low fibrinogen levels, which may obscure the analysis. In the absence of a generally accepted definition of excessive bleeding in cardiac surgery, we defined arbitrarily excessive bleeding as more than 1,000 mL/12 hours, which is twice the mean volume in the present study and involved 10% of the patients. Previous studies have used different definitions of excessive bleeding. Another potential limitation was that patients who underwent operations on the ascending aorta were excluded. In conclusion, this study found a preoperative plasma fibrinogen concentration of less than 2.5 g/L was associated with increased risk of excessive bleeding after cardiac operations. Fibrinogen concentration did not influence RBC transfusion requirements. This work was supported by V€astra G€ otaland region, the Gothenburg Medical Society, and the Swedish Heart and Lung Foundation. The sponsors of the study had no influence on the analysis and interpretation of data, in the writing of the report, or in the decision to submit the paper for publication.
References 1. Josefy S, Briones R, Bryant BJ. Preoperative coagulation studies to predict blood component usage in coronary artery bypass graft surgery. Immunohematology 2011;27:151–3.
ADULT CARDIAC
Ann Thorac Surg 2014;97:1199–206
1206
ET AL WALDEN FIBRINOGEN AND POSTOPERATIVE BLEEDING
ADULT CARDIAC
2. Karlsson M, Ternstrom L, Hyllner M, Baghaei F, Nilsson S, Jeppsson A. Plasma fibrinogen level, bleeding, and transfusion after on-pump coronary artery bypass grafting surgery: a prospective observational study. Transfusion 2008;48:2152–8. 3. Aljassim O, Karlsson M, Wiklund L, Jeppsson A, Olsson P, Berglin E. Inflammatory response and platelet activation after off-pump coronary artery bypass surgery. Scand Cardiovasc J 2006;40:43–8. 4. Blome M, Isgro F, Kiessling AH, et al. Relationship between factor XIII activity, fibrinogen, haemostasis screening tests and postoperative bleeding in cardiopulmonary bypass surgery. Thromb Haemost 2005;93:1101–7. 5. Gravlee GP, Arora S, Lavender SW, et al. Predictive value of blood clotting tests in cardiac surgical patients. Ann Thorac Surg 1994;58:216–21. 6. Wahba A, Rothe G, Lodes H, Barlage S, Schmitz G, Birnbaum DE. Predictors of blood loss after coronary artery bypass grafting. J Cardiothorac Vasc Anesth 1997;11:824–7. 7. Liu G, McNicol PL, McCall PR, et al. Prediction of the mediastinal drainage after coronary artery bypass surgery. Anaesth Intensive Care 2000;28:420–6. 8. Bolliger D, Gonsahn M, Levy JH, Williams WH, Tanaka KA. Is preoperative fibrinogen predictive for postoperative bleeding after coronary artery bypass grafting surgery? Transfusion 2009;49:2006–7. 9. Ternstrom L, Radulovic V, Karlsson M, et al. Plasma activity of individual coagulation factors, hemodilution and blood loss after cardiac surgery: a prospective observational study. Thromb Res 2010;126:e128–33. 10. Emeklibas N, Kammerer I, Bach J, Sack FU, Hellstern P. Preoperative hemostasis and its association with bleeding and blood component transfusion requirements in cardiopulmonary bypass surgery. Transfusion 2013;53:1226–34. 11. Poston R, Gu J, Manchio J, et al. Platelet function tests predict bleeding and thrombotic events after off-pump coronary bypass grafting. Eur J Cardiothorac Surg 2005;27:584–91. 12. Clauss A. [Rapid physiological coagulation method in determination of fibrinogen]. Acta Haematol 1957;17:237–46.
Ann Thorac Surg 2014;97:1199–206
13. Cockcroft DW, Gault MH. Prediction of creatinine clearance from serum creatinine. Nephron 1976;16:31–41. 14. The Fibrinogen Studies Collaboration. Associations of plasma fibrinogen levels with established cardiovascular disease risk factors, inflammatory markers, and other characteristics: Individual participant meta-analysis of 154,211 adults in 31 prospective studies. Am J Epidemiol 2007;166: 867–79. 15. Hellstern P, Muntean W, Schramm W, Seifried E, Solheim BG. Practical guidelines for the clinical use of plasma. Thromb Res 2002;107:S53–7. 16. Stainsby D, MacLennan S, Thomas D, Isaac J, Hamilton PJ. Guidelines on the management of massive blood loss. Br J Haematol 2006;135:634–41. 17. Carling MS, Jeppsson A, Wessberg P, Henriksson A, Baghaei F, Brisby H. Preoperative fibrinogen plasma concentration is associated with perioperative bleeding and transfusion requirements in scoliosis surgery. Spine 2011;36: 549–55. 18. Karlsson M, Ternstrom L, Hyllner M, et al. Prophylactic fibrinogen infusion reduces bleeding after coronary artery bypass surgery. A prospective randomised pilot study. Thromb Haemost 2009;102:137–44. 19. Fenger-Eriksen C, Lindberg-Larsen M, Christensen AQ, Ingerslev J, Sorensen B. Fibrinogen concentrate substitution therapy in patients with massive haemorrhage and low plasma fibrinogen concentrations. Br J Anaesth 2008;101: 769–73. 20. Rahe-Meyer N, Solomon C, Hanke A, et al. Effects of fibrinogen concentrate as first-line therapy during major aortic replacement surgery: a randomized, placebocontrolled trial. Anesthesiology 2013;118:40–50. 21. Buddeberg F, Schimmer BB, Spahn DR. Transfusion-transmissible infections and transfusion-related immunomodulation. Best Pract Res Clin Anaesthesiol 2008;22:503–17. 22. Prohaska W, Zittermann A, Inoue K, et al. Preoperative haemostasis testing does not predict requirement of blood products in cardiac surgery. Eur J Med Res 2008;13:525–30.