Bleeding occurred in 60% of all patients receiving vaECMO support during the course of therapy, whereas 80% of all patients with vvECMO support suffered from a bleeding event (Table 4). The most common locations were catheter and cannulation sites (70%) followed by bleeding from surgical wounds (27%), as well as diffuse dermal or mucosal bleeding (17%). Pulmonary bleeding occurred in 10 patients (9%), all of whom were treated with vvECMO. Retroperitoneal or central nervous system hemorrhage was rarely observed, each concerning only a single patient. Considering the number of days on which ECMO support was provided, bleeding occurred on 212 out of 530 days (40%).
Unfractioned heparin was used in all but two patients for anticoagulation. In two patients, heparin was discontinued because of the suspicion of heparin-induced thrombocytopenia and replaced with argatroban. The unfractioned heparin dosage was adjusted from a median of 800 IE/h to 600 IE/h during bleeding. Partial thromboplastin time (PTTmax) was 58 (48–65) s at 600 (350–1,100) IE/h heparin on days of a bleeding event and 52 (44–66) s at 800 (500–1,050) IE/h heparin on days without bleeding, which was not significantly different. The same applied if heparin was stopped on days with a bleeding event.
Red blood cell units were given most frequently followed by the administration of platelets. On average, one red blood cell concentrate was transfused per day. Both red blood cells and platelets were given more frequently on days with a bleeding event. Lower platelet counts were accepted in the absence of bleeding. The minimal platelet count was 56,000/μl (39–81.5) on days with a bleeding event versus 44,000/μl (26–59.5) on days without a bleeding event (p = 0.038). Overall, platelet counts decreased from a median of 145,000/μl at the start of ECMO therapy to 104,000/μl on day seven. The administration of coagulation factor concentrates was nearly exclusively done on days with a bleeding event, as was the use of single factor concentrates. Recombinant factor VIIa was used as a rescue measure in four patients; factor VIII was administered in two patients. Table 5 lists all blood products, including specific coagulation factor concentrates administered over the study period.
Table 6 depicts the therapeutic measures against bleeding during ECMO therapy, as well as their respective success rate. Overall, bleeding could be stopped in 34% of the events. In this regard, the termination of ECMO support was the most successful intervention (93% success rate), followed by the administration of tranexamic acid (88% success rate), replacement of the oxygenator/ECMO system (83% success rate), the administration of factor VIIa (75% success rate), as well as repositioning of at least one ECMO cannula (75% success rate). Moreover, the administration of a median of 36 μg desmopressin lead to promising results with a success rate of 50%. Both desmopressin and tranexamic acid were only administered on days with a bleeding event. Interestingly, if no intervention was carried out, bleeding still stopped with a success rate of 31%. Dosage reductions of unfractioned heparin helped to stop bleeding in 37%, whereas the termination of heparin was an unsuccessful therapeutic mean.
The analysis of factors predisposing to bleeding during ECMO therapy requiring therapeutic intervention revealed a number of risk factors (see Table S1, Supplemental Digital Content, http://links.lww.com/ASAIO/A136). These included a body weight below 83 kg (p = 0.002), fungal pneumonia (p < 0.001), interstitial lung disease (p = 0.044), and a lower peak inspiratory pressure (p = 0.047). The overall severity of disease did not increase the risk of bleeding. Moreover, there was no correlation between the occurrence of bleeding events and the survival of the patients.
The four variables that tested significant in the univariate log-rank analysis were included in a Cox regression model. In this multivariate model, only fungal pneumonia remained as a significant factor for the occurrence of a bleeding event requiring therapeutic intervention (p = 0.031; Table 7).
Management of hemostasis and anticoagulation represents a substantial challenge during ECMO. Bleeding complications are frequent and act as an important factor not only limiting therapeutic efficiency, but also survival of the patients.5 In the current retrospective analysis, we found that the majority of ECMO patients suffered from a bleeding event during the course of therapy.
Bleeding from the cannulation sites was most frequent, whereas major bleeding events were rarely observed. Termination of ECMO support was the most successful intervention; however, the administration of tranexamic acid, the repositioning of at least one ECMO cannula, or the replacement of the oxygenator/ECMO system yielded similar success rates. Considering important baseline characteristics, especially fungal pneumonia acted as a risk factor of bleeding. The specific mode of ECMO therapy or the severity of disease did not alter the risk or outcome of bleeding.
We included all bleeding events that had been documented within the patient management system without discriminating between major or minor hemorrhage. In this regard, 60–80% of the patients actually suffered from bleeding during the course of therapy. This incidence is higher than previously reported values of 26–54%, mainly because of the use of different study protocols and heterogeneous definitions of bleeding.7–9 In particular, a retrospective analysis complicates the precise assessment of the severity of bleeding, as therapeutic actions are usually taken at an early stage, for example, limiting the informative value of laboratory parameters. Moreover, the exact blood loss is difficult to determine and often not accurately documented.
Hence, the analysis of all documented bleeding events, as well as their high incidence, best reflects the overall impact of the coagulopathy onto the treatment of ECMO patients. Subsequently, analyzing the time to the first bleeding event requiring a therapeutic intervention appeared to be most reasonable for the survival curve statistics. Although major bleeding events as defined by current ELSO guidelines10 were rarely observed, one third of the patients still required at least one red blood cell unit per day. Even higher needs have been reported in previous studies,11,12 most likely because transfusion requirements not only depend on the incidence and severity of bleeding but also on the transfusion threshold. This varies between different ECMO centers. Although the ELSO suggests a transfusion threshold of 12–14 g/dl,10,13 the optimal hemoglobin concentration during ECMO therapy remains controversial. A restrictive use of red blood cells, targeting a hemoglobin concentration of 7.0 g/dl, has recently been shown to result in fewer bleeding complications with similar survival rates.7 In our institution, we usually use a transfusion threshold of 10 g/dl of hemoglobin during ECMO support. Nevertheless, it is important to consider that the liberal use of red blood cell concentrates has been associated with worse outcomes in critically ill patients at least in the absence of cardiac disease and a similar degree of organ dysfunction.14,15 In ECMO patients, this interrelation has not been established, and the complexity of the coagulopathy, which includes both pro- and anticoagulation, complicates the development of treatment guidelines during extracorporeal support.
Interestingly, anticoagulation management was not a major factor in the current analysis as dosage reductions or the termination of unfractioned heparin only lead to minor, nonsignificant reductions in the PTT. Stopping anticoagulation was an unsuccessful therapeutic mean. Furthermore, the observed success rate of 37% in terminating bleeding events after the reduction of the heparin dosage was contrasted by a 31% success rate when no intervention was carried out. However, the targeted PTT between 50 and 60 s may be considered a blood conserving measure in the first place,7 and the plethora of factors impacting the risk of bleeding complicates the determination of an optimal level of anticoagulation. In particular, increasing evidence suggests that platelet consumption and platelet dysfunction are hallmarks of the ECMO-induced coagulopathy. An acquired von Willebrand syndrome is common during extracorporeal support because of the loss of high molecular weight von Willebrand factor multimers.16–18 As such, the transfusion of platelets was a successful mean against bleeding. Bleeding events resolved in half of the patients; moreover, a similar success rate was achieved via the utilization of desmopressin to enhance platelet function. Platelets were more frequently transfused in the presence of hemorrhage, highlighting an increased awareness of this problem.
Moreover, the identification and relative importance of risk factors of bleeding during ECMO remains a pivotal task. A prior study in adults revealed that the use of centrifugal compared with roller pumps was associated with a higher risk of bleeding.19 A recent retrospective analysis of the ELSO Registry showed that mediastinal exploration, greater surgical complexity, early postoperative cannulation, as well as longer bypass times were important risk factors during pediatric ECMO. The role of systemic anticoagulation remained unclear.20 In our retrospective analysis, we were able to identify a number of risk factors of bleeding during adult ECMO. Lower body weight, fungal pneumonia, structural lung disease, and a low peak inspiratory pressure elevated the risk of bleeding in the univariate analysis; however, only fungal pneumonia prompted therapeutic interventions in the multivariate Cox regression model. Fungal infections are a rare cause of pneumonia accompanied by risk factors, such as immunosuppression, malignancy, organ transplant, or chronic steroid use. Patients with fungal pneumonia often suffer from, for example, concurring acute leukemia or lymphoma and myelosuppression during myeloablative chemotherapy. Fungal pneumonia has been previously identified to increase the odds of mortality during ECMO support for pediatric respiratory failure by approximately sixfold.21 Clinical courses are often complicated by multiorgan failure. Although data are scarce, this likely includes disseminated intravascular coagulation. In a small study, Lai et al.22 reported a rate of 58% during invasive pulmonary aspergillosis, whereas disseminated intravascular coagulation was the only factor directly associated with death. The association of coagulopathy with invasive pulmonary aspergillosis has been described early on, whereas consumption coagulopathy may be triggered in the vasculature via heavy infiltration with fungal hyphae. The possibility that the coagulopathy may be because of proteolytic enzymes produced by Aspergillus has also been suggested.23 Renal failure and dialysis had previously been shown to predispose to intracranial hemorrhage during ECMO support along with female gender and thrombocytopenia,24 suggesting a connection between an increased risk of bleeding and complicated treatment conditions.
We obtained inconsistent data regarding body weight and the risk of bleeding. Although a body weight below 83 kg was associated with a higher incidence of bleeding, the body mass index did not impact the occurrence of bleeding events. These findings may be explained by a higher volume ratio of the extracorporeal to the native circulation in patients with a lower body weight, which in turn may aggravate the coagulopathy and significantly increase the risk of bleeding. However, Al-Soufi et al.25 also showed that increased body weight is not a risk factor for in-hospital mortality in adult vvECMO.
The termination of ECMO support was the most successful mean to stop bleeding. This finding is not surprising as the pathophysiology of the acquired coagulopathy is complex and can probably only be entirely addressed by stopping the extracorporeal support. Recently, it has been shown that thrombocytopenia, factor XIII, and fibrinogen deficiency, as well as the loss of von Willebrand factor multimers all happen during ongoing ECMO support.18,26 As the replacement of the oxygenator yielded similar therapeutic success rates, platelet and fibrin deposition as well as the formation of small clots and fibrin strands adhering to the oxygenator and tubing may very well fuel the disseminated intravascular coagulation and generalized hyperfibrinolysis. As such, close monitoring of clinical and laboratory parameters including D-dimers must be emphasized.27,28 However, the exchange of the extracorporeal circuit might be a useful alternative in case stopping the ECMO support is not an option. Nevertheless, when assessing the usefulness of therapeutic interventions against bleeding, it is equally important to consider the different types and locations of the bleeding events. Although we could not subclassify the severity of hemorrhage according to a standardized bleeding definition in our retrospective analysis, the majority of the observed bleeding events certainly conformed to type 2 bleeding of the Bleeding Academic Research Consortium Definition for Bleeding.29 However, they did not conform to major bleeding events with hemoglobin drops of ≥3 to <5 g/d. These were rarely observed. Cannulation sites were the most frequent locations of bleeding, overall suggesting that diffuse blood loss was the greatest threat. This is in line with data from the ELSO registry also reporting cannulation sites as prime bleeding locations.30 As such, future studies should aim at a better characterization of the pathophysiology of the coagulation cascade, as well as the efficiency of antihemorrhagic agents. In this regard, the use of tranexamic acid and desmopressin provided good results in our study. The administration of recombinant factor VII was also very successful; nevertheless, its use was limited to life-threatening refractory hemorrhage, where it has been shown to provide an acceptable safety profile.31 Moreover, both tranexamic acid, desmopressin, or factor XIII were used more frequently towards the end of the study period (results not shown), suggesting an increasing need to counteract the coagulopathy during sustained ECMO support.
The current results must be interpreted within the constraints of several potential limitations. The study was conducted as a single-center study in a German ECMO center, and limited external validity cannot be excluded. This design was chosen to maximize standardized treatment, allow data analysis via a single patient data management system, as well as ease communication within the investigator and the ECMO team. However, our data represent the bleeding complications of an experienced ECMO center providing both vaECMO and vvECMO support, including cardiac and respiratory failure as well as extracorporeal cardiopulmonary resuscitation. Furthermore, the retrospective study design does not allow randomization or blinding, and the analysis depends on the quality of the written record. In our institution, patient data are recorded with a standardized patient data management system, guaranteeing a high level of detail and comparability within the documentation with regard to baseline characteristics, laboratory values, imaging, ventilation, ECMO parameters, or therapeutic interventions, respectively. However, the magnitude of bleeding events was not recorded according to a standardized bleeding definition, as such, the documentation of bleeding events and their magnitude depends on an individual case-by-case evaluation. Risk factors of bleeding were analyzed with univariate testing, and interdependencies of the parameters cannot be fully excluded. Multivariate testing was performed with factors that tested significant in univariate analysis. However, statistical power was low according to the limited number of bleeding events per factor introduced into the model. In addition, some interrelated factors may be missed.
In conclusion, the current single-center study shows that bleeding during ECMO remains a major diagnostic and therapeutic challenge. In this retrospective analysis, 60–80% of all patients suffered from bleeding during the course of ECMO support. Fungal pneumonia was the only identifiable risk factor, whereas the termination of ECMO was the most successful therapeutic intervention. The use of tranexamic acid and desmopressin provided promising results. Although our data are limited by the aforementioned constraints, they represent the bleeding complications and corresponding therapeutic interventions of an experienced ECMO center, as such providing an important basis for future randomized trials.
1. Brodie D, Bacchetta M. Extracorporeal membrane oxygenation for ARDS in adults. N Engl J Med 2011.365: 1905–1914.
2. Oliver WC. Anticoagulation and coagulation management for ECMO
. Semin Cardiothorac Vasc Anesth 2009.13: 154–175.
3. Kumar TK, Zurakowski D, Dalton H, et al. Extracorporeal membrane oxygenation in postcardiotomy patients: factors influencing outcome. J Thorac Cardiovasc Surg 2010, 140(2):330–336 e332.
4. Marasco SF, Vale M, Pellegrino V, et al. Extracorporeal membrane oxygenation in primary graft failure after heart transplantation. Ann Thorac Surg 2010.90: 1541–1546.
5. Aubron C, Cheng AC, Pilcher D, et al. Factors associated with outcomes of patients on extracorporeal membrane oxygenation support: a 5-year cohort study. Crit Care 2013.17: R73.
6. Lamb KM, Cowan SW, Evans N, et al. Successful management of bleeding
complications in patients supported with extracorporeal membrane oxygenation with primary respiratory failure. Perfusion 2013.28: 125–131.
7. Agerstrand CL, Burkart KM, Abrams DC, Bacchetta MD, Brodie D. Blood conservation in extracorporeal membrane oxygenation for acute respiratory distress syndrome. Ann Thorac Surg 2015.99: 590–595.
8. Davies A, Jones D, Bailey M, et al; Australia and New Zealand Extracorporeal Membrane Oxygenation (ANZ ECMO
) Influenza Investigators: Extracorporeal Membrane Oxygenation for 2009 Influenza A(H1N1) Acute Respiratory Distress Syndrome. JAMA 2009.302: 1888–1895.
9. Patroniti N, Zangrillo A, Pappalardo F, et al. The Italian ECMO
network experience during the 2009 influenza A(H1N1) pandemic: preparation for severe respiratory emergency outbreaks. Intensive care medicine 2011, 37(9):1447–1457.
11. Butch SH, Knafl P, Oberman HA, Bartlett RH. Blood utilization in adult patients undergoing extracorporeal membrane oxygenated therapy. Transfusion 1996.36: 61–63.
12. Norfolk SG, Hollingsworth CL, Wolfe CR, et al. Rescue therapy in adult and pediatric patients with pH1N1 influenza infection: a tertiary center intensive care unit experience from April to October 2009. Critical care medicine 2010, 38(11):2103–2107.
13. Annich GM. Extracorporeal life support: the precarious balance of hemostasis. J Thromb Haemost 2015.13(suppl 1): S336–S342.
14. Hébert PC, Wells G, Blajchman MA, et al. A multicenter, randomized, controlled clinical trial of transfusion requirements in critical care. Transfusion Requirements in Critical Care Investigators, Canadian Critical Care Trials Group. N Engl J Med 1999.340: 409–417.
15. Vincent JL, Baron JF, Reinhart K, et al; ABC (Anemia and Blood Transfusion in Critical Care) Investigators: Anemia and blood transfusion in critically ill patients. JAMA 2002.288: 1499–1507.
16. Heilmann C, Geisen U, Beyersdorf F, et al. Acquired von Willebrand syndrome in patients with extracorporeal life support (ECLS). Intensive Care Med 2012.38: 62–68.
17. Kalbhenn J, Schmidt R, Nakamura L, Schelling J, Rosenfelder S, Zieger B. Early diagnosis of acquired von Willebrand Syndrome (AVWS) is elementary for clinical practice in patients treated with ECMO
therapy. J Atheroscler Thromb 2015.22: 265–271.
18. Tauber H, Ott H, Streif W, et al. Extracorporeal membrane oxygenation induces short-term loss of high-molecular-weight von Willebrand factor multimers. Anesth Analg 2015.120: 730–736.
19. Halaweish I, Cole A, Cooley E, Lynch WR, Haft JW. Roller and centrifugal pumps: a retrospective comparison of bleeding
complications in extracorporeal membrane oxygenation. ASAIO J 2015.61: 496–501.
20. Werho DK, Pasquali SK, Yu S, et al; Extracorporeal Life Support Organization Member Centers: Hemorrhagic complications in pediatric cardiac patients on extracorporeal membrane oxygenation: an analysis of the Extracorporeal Life Support Organization Registry. Pediatr Crit Care Med 2015.16: 276–288.
21. Zabrocki LA, Brogan TV, Statler KD, Poss WB, Rollins MD, Bratton SL. Extracorporeal membrane oxygenation for pediatric respiratory failure: Survival and predictors of mortality. Crit Care Med 2011.39: 364–370.
22. Lai CC, Liaw SJ, Lee LN, Hsiao CH, Yu CJ, Hsueh PR. Invasive pulmonary aspergillosis: high incidence of disseminated intravascular coagulation in fatal cases. J Microbiol Immunol Infect 2007.40: 141–147.
23. McClellan SL, Komorowski RA, Farmer SG, Hussey CV, Kauffman HM Jr, Adams MB. Severe bleeding
diathesis associated with invasive aspergillosis in transplant patients. Transplantation 1985.39: 406–410.
24. Kasirajan V, Smedira NG, McCarthy JF, Casselman F, Boparai N, McCarthy PM. Risk factors for intracranial hemorrhage in adults on extracorporeal membrane oxygenation. Eur J Cardiothorac Surg 1999.15: 508–514.
25. Al-Soufi S, Buscher H, Nguyen ND, Rycus P, Nair P. Lack of association between body weight and mortality in patients on veno-venous extracorporeal membrane oxygenation. Intensive Care Med 2013.39: 1995–2002.
26. Kalbhenn J, Wittau N, Schmutz A, Zieger B, Schmidt R. Identification of acquired coagulation disorders and effects of target-controlled coagulation factor substitution on the incidence and severity of spontaneous intracranial bleeding
during veno-venous ECMO
therapy. Perfusion 2015.30: 675–682.
27. Lubnow M, Philipp A, Dornia C, et al. D-dimers as an early marker for oxygenator exchange in extracorporeal membrane oxygenation. J Crit Care 2014, 29(3):473 e471–475.
28. Lubnow M, Philipp A, Foltan M, et al. Technical complications during veno-venous extracorporeal membrane oxygenation and their relevance predicting a system-exchange–retrospective analysis of 265 cases. PLoS One 2014.9: e112316.
29. Mehran R, Rao SV, Bhatt DL, et al. Standardized bleeding
definitions for cardiovascular clinical trials: a consensus report from the Bleeding
Academic Research Consortium. Circulation 2011.123: 2736–2747.
30. Paden ML, Conrad SA, Rycus PT, Thiagarajan RR, Registry E. Extracorporeal Life Support Organization Registry Report 2012. ASAIO J 2013.59(3): 202–210.
31. Anselmi A, Guinet P, Ruggieri VG, et al. Safety of recombinant factor VIIa in patients under extracorporeal membrane oxygenation. Eur J Cardiothorac Surg 2016.49: 78–84.
Supplemental Digital Content
Copyright © 2017 by the American Society for Artificial Internal Organs