Secondary Logo

Journal Logo

Clinical Critical Care

Severe Thrombocytopenia in Adults with Severe Acute Respiratory Distress Syndrome: Impact of Extracorporeal Membrane Oxygenation Use

Dzierba, Amy L.*; Roberts, Russel†‡; Muir, Justin*; Alhammad, Abdullah§; Schumaker, Greg; Clark, Jacqueline; Ruthazer, Robin#; Devlin, John W.†¶

Author Information
doi: 10.1097/MAT.0000000000000415


Technological improvements associated with extracorporeal membrane oxygenation (ECMO) over the past decade have resulted in the increased use and the expansion of new centers offering this supportive care.1 Specifically, venovenous ECMO has been used to manage patients suffering from severe hypoxemic respiratory failure, because of illnesses like acute respiratory distress syndrome (ARDS), who fail to improve with traditional management strategies.2–6

Thrombocytopenia frequently occurs in critically ill patients.7–9 When patients with severe ARDS experience thrombocytopenia, complications such as bleeding and thrombosis may occur that can truncate or preclude ECMO use.1 If thrombocytopenia becomes severe, transfusion is often required, the risk for a serious bleeding event escalates, and the likelihood of intensive care unit (ICU) mortality increases.9–11 The etiology of thrombocytopenia in the critically ill patients is often hard to discern given that causes for thrombocytopenia are numerous and the processes that govern platelet destruction, sequestration, production, and hemodilution remain poorly understood.10,12

Small case series and reviews suggest the extracorporeal circuit used in ECMO may cause thrombocytopenia through a process of non-immune-mediated enhanced platelet consumption.8,13–16 In contrast, a recent retrospective cohort study of 100 adults who received ECMO for acute respiratory failure and that controlled for severity of illness, platelet transfusion use and administration of potential thrombocytopenia-inducing medications, found that the duration of ECMO use was not associated with the development of thrombocytopenia.17 However, with thrombocytopenia prevalent in critically ill patients not managed with ECMO, the failure of these authors to include a non-ECMO control group in their analysis precludes the ability to know whether ECMO use is an independent risk factor for thrombocytopenia. The objective of this study was to characterize the incidence and factors associated with severe thrombocytopenia among adults with severe ARDS managed with or without ECMO.


This retrospective cohort analysis included consecutive adults with severe ARDS (on at least 1 day of ICU stay) over a 5-year period (2009–2013) managed with ECMO at (NewYork-Presbyterian Hospital, Columbia University Medical Center [NYP/CUMC] in New York, NY) or without ECMO at Tufts Medical Center in Boston, MA. At NYP/CUMC, the ECMO circuit consists of either the combination of a Rotaflow centrifugal pump (Maquet, Rastatt, Germany), Quadrox I or D oxygenator (Maquet, Rastatt, Germany), or the CARDIOHELP support system (Maquet, Rastatt, Germany). Approval for this study was obtained from the Institutional Review Boards at both institutions.

Severe ARDS was defined using the Berlin criteria (i.e., a ratio of partial pressure of oxygen in arterial blood to fraction of inspired oxygen [PaO2:FiO2] ≤ 100 with a positive end expiratory pressure ≥ 5 cm H2O, new or worsening respiratory symptoms of less than one week in duration and bilateral opacities on chest radiograph not explained by cardiac failure or fluid overload).18 Patients with a history of chemotherapy in the prior 14 days, having severe thrombocytopenia (platelet count of ≤ 50,000/μl) at the time of severe ARDS onset, or who died within 48 hours of severe ARDS onset were excluded given the potential confounding nature of these factors on the planned analysis.

Data collected at ICU baseline included age, APACHE II score,19 and the presence of underlying medical condition(s) reported to be associated with a reduced platelet count.8,20–22 Data collected at both ICU baseline and each ICU day (until mechanical ventilation liberation ≥ 24 hours, ICU discharge or death [whatever came first]) included the lowest platelet count recorded, presence of liver failure (i.e., international normalized ratio ≥ 2 and a total bilirubin ≥ 1.5 mg/dl in a patient not receiving warfarin), presence of septic shock (as defined by the American College of Chest Physicians/Society of Critical Care Medicine definition and the administration of any dose of a vasopressor),23 presence of a serum lactate dehydrogenase (LDH) ≥ 300 unit/L, administration of UFH as a continuous infusion, prophylactic low-dose UFH, or a non-heparin medication(s) reported in the literature to be associated with an incidence of thrombocytopenia ≥ 1% and a transfusion of ≥1 unit of platelets. These medical conditions and non-heparin medications are listed in supplementary table (Supplemental Digital Content 1,

The development of severe thrombocytopenia after the onset of severe ARDS was the primary clinical outcome evaluated. This end-point was chosen given that it represents a platelet count that should be of clinical concern to the ICU team and, depending on other clinical factors present on the day it occurred, might increase the chance that an intervention such as a platelet transfusion would be administered. In an effort to avoid falsely attributing the presence of a potential risk factor for thrombocytopenia with the development of severe thrombocytopenia (or the platelet nadir during the ICU stay), the following risk factors were deemed to be present only when they preceded the severe thrombocytopenia onset (platelet nadir) by ≤2 days: UFH use, non-heparin medication associated with thrombocytopenia use, LDH elevation, and liver failure.

Baseline characteristics and potential thrombocytopenic factors were compared between the ECMO and non-ECMO groups. Subsequently, a univariate analysis compared potential thrombocytopenic factors, including ECMO use, between patients who developed severe thrombocytopenia with those who did not. All data are summarized as either percent, mean ± standard deviation (SD) or in the case of skewed distribution, as a median (interquartile range [IQR]). Comparisons between ECMO and non-ECMO groups were analyzed using a χ2 test, Student’s t-test, and the Kruskal–Wallis test where applicable. A random effects (mixed) generalized linear model, that accounted for the repeated (and sometime irregular nature) of platelet count evaluation, was used to estimate the platelet count over time for each treatment group. For the patients who developed severe thrombocytopenia, a survival analysis was used to model the time from severe ARDS onset to severe thrombocytopenia onset. A fully adjusted multivariable linear regression model was built using a stepwise selection process. The first variable in the model, and forced into subsequent models, was an indicator variable with a value of “1” to denote the presence of severe thrombocytopenia (and “0” to denote its absence). All variables that could affect the development of severe thrombocytopenia (if the unadjusted association had a p value < 0.01) and the use of ECMO were considered as candidates to step into the model. Unless otherwise stated, a p value ≤ 0.05 signified statistical significance. All analyses were done using SAS 9.2 for windows (SAS Institute, Cary, NC).


Extracorporeal Membrane Oxygenation Versus Nonextracorporeal Membrane Oxygenation Analysis

A total of 85 patients (32 ECMO and 53 non-ECMO) were included. Extracorporeal membrane oxygenation patients, when compared with non-ECMO patients, were younger, had fewer baseline medical conditions associated with a lower platelet count, were more likely to receive a continuous UFH infusion and to be administered non-heparin medications reported to be associated with thrombocytopenia (Table 1).

Table 1.
Table 1.:
Comparison of Baseline Characteristics and Potential Thrombocytopenia Factors Between the ECMO and Non-ECMO Groups

The change in platelet count over time between the ECMO and non-ECMO groups is presented in Figure 1. Whereas the daily platelet count over the course of the data collection period significantly increased in each group [ECMO (43,000/μl; p = 0.008) versus non-ECMO (47,000/μl; p = 0.02) groups], this degree of change was similar between the two groups (p = 0.98). The number of days from the onset of ARDS to the ICU day that the platelet count nadir was reached (as presented a median [IQR]) was similar between the ECMO (5 [2–16]) and non-ECMO (5 [2–15]) groups (p = 0.86).

Figure 1.
Figure 1.:
Change in the daily platelet count over the period of data collection between the ECMO and non-ECMO groups based on the use of a generalized linear model that accounts for repeated measures over time. Each line presents the trajectory of one patient with a black line indicating a patient managed with ECMO and a grey line indicating a patient not managed with ECMO. The fitted trajectories are plotted in bold where the solid black line indicates the average for ECMO patients and the dotted grey line indicates the average for non-ECMO patients. ECMO, extracorporeal membrane oxygenation.

Severe Thrombocytopenia Versus Nonsevere Thrombocytopenia Analysis

The proportion of patients who developed severe thrombocytopenia over the study period was similar between the ECMO and non-ECMO groups (25 vs. 19%, p = 0.5). Patients with severe thrombocytopenia were more likely to be administered ≥ 1 platelet transfusion (22 vs. 2%, p = 0.0009) but in no patient was a platelet transfusion administered before the platelet nadir being reached. Among these patients, the time to development of severe thrombocytopenia was similar between the two groups (p = 0.49) (Figure 2). The patients who developed severe thrombocytopenia (versus those that did not) while having a higher APACHE II score, an elevated LDH ≥ 300 unit/L and a trend for more liver failure (p = 0.08), were not more likely to have been managed with ECMO (p = 0.5; Table 2).

Table 2.
Table 2.:
Comparison of Potential Thrombocytopenia Factors Between Patients with and without Severe Thrombocytopenia
Figure 2.
Figure 2.:
Comparison of time to severe thrombocytopenia between patients in the ECMO (n = 32) and non-ECMO groups (n = 53). The time to severe thrombocytopenia was similar between the two groups (log-rank test comparing curves was p = 0.49). The number of patients where data collection was still occurring is plotted at the bottom of the figure in the box. Over the course of the period of data collection, the proportion of patients who ever developed severe thrombocytopenia was similar between the ECMO (8/32 [25%]) and non-ECMO (10/53 [19%]) groups (p = 0.5). ECMO, extracorporeal membrane oxygenation.

The fully adjusted multivariable linear regression model that was built on n = 85 subjects, where the presence of severe thrombocytopenia was the indicator variable, revealed that only baseline APACHE II score (p = 0.02) was independently associated with the presence of severe thrombocytopenia across the ECMO and non-ECMO groups, but the use of ECMO was not (p = 0.32; Table 3).

Table 3.
Table 3.:
Estimated Difference in Severe Thrombocytopenia Incidence Between the ECMO (n = 32) and Non-ECMO (n = 53) Groups: Adjusted Multivariable Linear Regression Model (R2 = 0.141)


Our investigation is the first to use a non-ECMO control group to characterize the impact of venovenous ECMO use on the platelet count and the incidence of severe thrombocytopenia in critically ill adults. Among adults with severe ARDS, use of ECMO does not appear to affect the change in the daily platelet count over the course of the ICU stay nor the proportion of patients who develop severe thrombocytopenia. After controlling for those factors with the potential to cause thrombocytopenia that differed between severe ARDS patients who developed severe thrombocytopenia and those that did not, ECMO use was not shown to influence the incidence of severe thrombocytopenia.

Important differences exist between our study and the results of other published reports that describe ECMO use to be associated with platelet declines in critically ill adults with severe ARDS or other primary lung conditions.13,15,16 Unlike these other reports, our analysis included a non-ECMO control group and carefully considered and adjusted for the many baseline and daily factors that might lead to reductions in the platelet count over the course of the ICU admission. Moreover, by controlling for the fact that platelet counts are repeatedly (and sometimes unpredictably measured) over time by using a generalized estimating equation approach to plot their change over the course of the ICU admission, we observed a small increase in the platelet count from the time of severe ARDS onset that was similar between both ECMO and non-ECMO groups.

Our study builds on the results of the recent report by Abrams et al.17 that found the duration of ECMO use was not associated with the development of severe thrombocytopenia after adjustment for factors that might influence the platelet count like severity of illness and administration of potential thrombocytopenic medications. However, by not including a control group of similar patients not managed with ECMO, they were not able to consider whether ECMO use is independently associated with the development of severe thrombocytopenia.

Our study has a number of potential limitations. Its retrospective nature precludes the ability to understand why potential risk factors for thrombocytopenia were present and interventions implemented by clinicians when thrombocytopenia occurred. There are additional factors that could influence the incidence of thrombocytopenia that was reported that were not considered such as hemodilution, evolving sepsis not yet manifesting as vasopressor-dependent septic shock or ECMO circuit changes. Factors influencing the platelet count that were present before the diagnosis of severe ARDS or transport to NYP/CUMC may have influenced our results but were not able to be considered. A standard 2 day pre-nadir time frame was used for all thrombocytopenia risk factors despite the fact that the time to cause thrombocytopenia likely varies by the risk factor involved. Whereas factors associated with thrombocytopenia present two or more days before the platelet nadir was reached could have influenced the platelet nadir reported, they were not considered. With the exception of UFH, the dose of a potentially thrombocytopenic medication or the severity of potentially thrombocytopenic conditions were not considered.

Our analysis solely attributes clinical significance to the platelet count itself when other factors such as platelet functionality or non-thrombocytopenic risks for bleeding might also be just as important to ICU clinicians. Moreover, neither the presence nor severity of thrombocytopenia-associated bleeding was considered in our analysis given its retrospective nature. Very severe thrombocytopenia (i.e., platelets ≤ 25,000/μl) may be an outcome that is more clinically relevant to clinicians but it occurred at a frequency too low to be considered in the analysis. Use of a formal matching process to identify patients for each cohort was deemed unfeasible given the limited number of patients with severe ARDS who are managed with ECMO even at a high-volume center like NYP. Our results may be different at other institutions where clinical practices are different or if patients who died within 48 hours of severe ARDS onset had been included. The small study sample size and the low coefficient of determination (R2) generated by the regression model each limit the conclusions that can be made surrounding the effect of ECMO use on severe thrombocytopenia occurrence in patients with severe ARDs. Finally, the ability to extrapolate our results to non-ARDS patients where ECMO may also be used may be limited.

Critically ill adults receiving ECMO represent a complex group of patients who frequently have multiple concomitant factors associated with thrombocytopenia. Whereas our analysis suggests that ECMO use may not be independently associated with either a reduction in the platelet account or a risk factor for severe thrombocytopenia, the size and design of our analysis precludes the ability to make strong recommendations from its results. Clinicians should continue to identify and remove (where possible) non-EMCO causes for thrombocytopenia when it occurs. Future research, that is both prospective and well-controlled, includes larger patient populations, and carefully evaluates the temporal relationship between thrombocytopenia risk factors and thrombocytopenia development, are needed to further investigate the relationship between ECMO use and thrombocytopenia in critically ill adults.


1. Paden ML, Conrad SA, Rycus PT, Thiagarajan RRExtracorporeal life support organization registry report 2012.ASAIO J201359202210
2. Davies A, Jones D, Bailey M, et al.Extracorporeal membrane oxygenation for 2009 influenza A (H1N1) acute respiratory distress syndrome.JAMA200930218881895
3. Peek GJ, Mugford M, Tiruvoipati R, et al.CESAR Trial CollaborationEfficacy and economic assessment of conventional ventilatory support versus extracorporeal membrane oxygenation for severe adult respiratory failure (CESAR): a multicentre randomized controlled trial.Lancet200937413511363
4. Noah MA, Peek GJ, Finney SJ, et al.Referral to an extracorporeal membrane oxygenation center and mortality among patients with severe 2009 influenza A (H1N1).JAMA201130616591668
5. Michaels AJ, Hill JG, Long WB, et al.Adult refractory hypoxemic acute respiratory distress syndrome treated with extracorporeal membrane oxygenation: the role of a regional referral center.Am J Surg2013205492498
6. Combes A, Brodie D, Bartlett R, et al.Position paper for the organization of extracorporeal membrane oxygenation programs for acute respiratory failure in adult patients.Am J Respir Crit Care Med2014190488496
7. Bonfiglio MF, Traeger SM, Kier KL, et al.Thrombocytopenia in intensive care patients: a comprehensive analysis of risk factors in 314 patients.Ann Pharmacother199529835842
8. Greinacher A, Selleng KThrombocytopenia in the intensive care unit patient.Hematology Am Soc Hematol Educ Program2010354351
9. Crowther MA, Cook DJ, Meade MO, et al.Thrombocytopenia in medical-surgical critically ill patients: prevalence, incidence, and risk factors.J Crit Care200520348353
10. Baughman RP, Lower EE, Flessa HC, Tollerud DJThrombocytopenia in the intensive care unit.Chest199310412431247
11. Vanderschueren S, De Weerdt A, Malbrain M, et al.Thrombocytopenia and prognosis in intensive care.Crit Care Med20002818711876
12. Lequier L, Annich G, Al-Ibrahim O, et al.ELSO Anticoagulation Guidelines2013The Extracorporeal Life Support Organization (ELSO)117
13. Glick D, Dzierba AL, Abrams D, et al.Clinically suspected heparin-induced thrombocytopenia during extracorporeal membrane oxygenation.J Crit Care20153011901194
14. Cheung PY, Sawicki G, Salas E, Etches PC, Schulz R, Radomski MWThe mechanisms of platelet dysfunction during extracorporeal membrane oxygenation in critically ill neonates.Crit Care Med20002825842590
15. Panigada M, Artoni A, Passamonti SM, et al.Hemostasis changes during veno-venous extracorporeal membrane oxygenation for respiratory support in adults.Minerva Anestesiol201682170179
16. Weingart C, Lubnow M, Philipp A, et al.Comparison of coagulation parameters, anticoagulation, and need for transfusion in patients on interventional lung assist or veno-venous extracorporeal membrane oxygenation.Artif Organs201539765773
17. Abrams D, Baldwin MR, Champion M, et al.Thrombocytopenia and extracorporeal membrane oxygenation in adults with acute respiratory failure: a cohort study.Intensive Care MedEpub ahead of print
18. Ranieri VM, Rubenfeld GD, Thompson BT, et al.Acute respiratory distress syndrome: the Berlin definition.JAMA201230725262533
19. Knaus WA, Draper EA, Wagner DP, Zimmerman JEAPACHE II: a severity of disease classification system.Crit Care Med198513818829
20. Priziola JL, Smythe MA, Dager WEDrug-induced thrombocytopenia in critically ill patients.Crit Care Med2010386 supplS145S154
21. Reese JA, Li X, Hauben M, et al.Identifying drugs that cause acute thrombocytopenia: an analysis using 3 distinct methods.Blood201011621272133
22. Jones KL, Kiel PJTisdale JE, Miller DAChapter 47. Drug Induced Diseases20101st edBethesda, MDAmerican Society of Health System Pharmacistspp. 929940
23. Dellinger RP, Levy MM, Rhodes A, et al.Surviving sepsis campaign: international guidelines for management of severe sepsis and septic shock.Crit Care Med201241580637

platelet; acute respiratory distress syndrome; extracorporeal membrane oxygenation; intensive care unit

Supplemental Digital Content

Copyright © 2016 by the ASAIO