Abstract
Extracorporeal membrane oxygenation (ECMO) provides lifesaving circulatory support and gas exchange, although hematologic complications are frequent. The relationship between ECMO and severe thrombocytopenia (platelet count <50 × 109/L) remains ill defined. We performed a cohort study of 67 patients who received ECMO between 2016 and 2019, of which 65.7% received veno-arterial (VA) ECMO and 34.3% received veno-venous (VV) ECMO. All patients received heparin and 25.4% received antiplatelet therapy. In total, 23.9% of patients had a thrombotic event and 67.2% had a haemorrhagic event. 38.8% of patients developed severe thrombocytopenia. Severe thrombocytopenia was more common in patients with lower baseline platelet counts and increased the likelihood of thrombosis by 365% (OR 3.65, 95% CI 1.13–11.8, P=0.031), while the type of ECMO (VA or VV) was not predictive of severe thrombocytopenia (P=0.764). Multivariate logistic regression controlling for additional clinical variables found that severe thrombocytopenia predicted thrombosis (OR 3.65, CI 1.13–11.78, P=0.031). Over a quarter of patients requiring ECMO developed severe thrombocytopenia in our cohort, which was associated with an increased risk of thrombosis and in-hospital mortality. Additional prospective observation is required to clarify the clinical implications of severe thrombocytopenia in the ECMO patient population.
Keywords: ECMO, anticoagulation, thrombosis, thrombocytopenia
Introduction:
Extracorporeal membrane oxygenation (ECMO) can provide lifesaving circulatory support and gas exchange via a cardiopulmonary bypass circuit to critically ill and postsurgical patients; however, significant rates of hematologic complications compromise its use. There are two predominant types of ECMO: veno-arterial (VA) and veno-venous (VV). In VA ECMO, deoxygenated blood is directed through the ECMO circuit and an external oxygenator before returning to the body via arterial cannulation, thus replacing both oxygenation and circulatory support. In contrast, VV ECMO provides only gas exchange via VV cannulation techniques.
Thrombosis and hemorrhage are exceedingly common, with overall bleeding rates as high as 60% reported in patients on ECMO.1 The pathophysiology of ECMO-associated hemorrhage is complex, involving the mandated use of parenteral anticoagulation, consumption of coagulation factors, acquired deficiency of von Willebrand factor, platelet activation/dysfunction, and subsequent platelet consumption leading to thrombocytopenia.2,3 Paradoxically, thrombosis (both venous and arterial) is similarly common in the critically ill population requiring ECMO, with rates of cannula-associated deep vein thrombosis approaching 85%.4 Thus, ECMO results in a precarious balance between hemostasis and thrombosis, contributing to high rates of morbidity and mortality.
Thrombocytopenia is a well-described phenomenon associated with ECMO, with an estimated prevalence of 21%; severe thrombocytopenia (platelet count <50 × 109/L) in patients on ECMO, however, is not as well characterized and has a reported prevalence of 6.3–26.6% in the literature.5 While thrombocytopenia and severe thrombocytopenia have been associated with reduced survival in patients on ECMO, the relationship between severe thrombocytopenia and specific clinical outcomes such as bleeding and thrombosis has not been fully evaluated.5 Additionally, the contribution of platelet counts to the complex milieu of ECMO-associated bleeding and thrombosis is poorly understood.
We performed the following retrospective cohort study to better define the incidence, predictors, and clinical consequences of severe thrombocytopenia in patients on ECMO.
Methods
Study Design and Data Source
This study protocol was approved by the Oregon Health and Science University (OHSU) Institutional Review Board prior to study initiation (Study Number 00019765). We performed a retrospective cohort study of adults (age ≥18 years) who received ECMO (either VA or VV) for any indication at OHSU between March 1, 2016 and September 30, 2019. All data were obtained by extraction from the hospital’s electronic medical record.
Study Population
A total of 69 patients were treated with ECMO during the study period. Patients who had a documented ECMO course between 1 and 90 days were included, resulting in a final study population of 67 adult patients.
Data Collection
Data were extracted by direct medical record review including patient demographics, laboratory parameters, clinical indication for ECMO use, duration of treatment, ECMO parameters, anticoagulant and antiplatelet use and monitoring parameters (partial thromboplastin time (aPTT) and anti-Factor Xa assay (anti-FXa), and safety. Clinical outcomes assessed included overall survival, rate of hemorrhage, thrombosis, and blood transfusion requirements while on ECMO. Hemorrhage was defined by the International Society on Thrombosis and Haemostasis (ISTH) consensus criteria.6 Both venous and arterial thrombosis were included, as well as thrombotic events within the extracorporeal circuit that mandated a circuit exchange. Asymptomatic patients on ECMO are not routinely screened for venous thrombosis in our hospital system. Patients generally only undergo diagnostic evaluation based on clinical suspicion for thrombosis. The first 30 complete blood counts obtained while on ECMO were evaluated in order to determine the incidence of thrombocytopenia and platelet trends over time. On average, there were 3.7 platelet checks per day in the cohort. The average duration of ECMO treatment in the cohort was 7.0 ± 5.0 days.
ECMO Anticoagulation Protocols
In the absence of a major contradiction (bleeding within a critical site or life-threatening bleeding), heparin is administered in conjunction with ECMO to achieve a therapeutic anti-FXa level of 0.35–0.70 IU/mL for VA ECMO and 0.30–0.50 IU/mL for VV ECMO. Antiplatelet therapy and cannulation strategy were determined at the discretion of the treating physicians. The majority of patients received ECMO via the Maquet Cardiohelp life support system.
Statistical Analysis
Patient demographics were reported in total and by type of ECMO (VA vs. VV) using standard descriptive statistics. Continuous data were presented as a mean value with standard deviation (SD). Continuous and dichotomous variables were compared between VV and VA ECMO with Welch’s t-tests and Fisher exact tests, respectively. Statistical significance was defined as a two-sided P value of <0.05 for all analyses.
Univariate logistic regression analysis was used to identify risk factors for the following specific clinical outcomes of interest; the presence of documented bleeding events, major hemorrhage by ISTH criteria,6 and thrombosis. The following variables were selected apriori to be assessed in the model: (i) mean platelet count, (ii) platelet count range, or (iii) severe thrombocytopenia (platelet count <50 × 109/L).
Multiple logistic regression analysis was then used to identify risk factors for the aforementioned outcomes, and for the development of severe thrombocytopenia. The following variables were selected apriori and considered as independent predictors in the model: (i) age, (ii) concurrent use of antiplatelet drugs, (iii) duration of ECMO, (iv) body mass index (BMI), (v) severe thrombocytopenia, and (vi) mean albumin level.
All statistics were calculated using R (R Foundation for Statistical Computing, version 3.6.2).
Results
Patient Demographics and Clinical Outcomes
During the 43-month study period, 67 adult patients treated with ECMO met inclusion criteria. Within this cohort, 44 (65.7%) received VA ECMO and 23 (34.3%) received VV ECMO. The cohort had a mean age of 51.1 years and 65.7% of the entire study population identified as male. The mean duration of ECMO was 7.0 days. The mean platelet count while on ECMO was 118 × 109/L. However, patients who went on to develop severe thrombocytopenia had lower baseline platelet counts (199 × 109/L vs. 110 × 109/L, P<0.001). In total, only 58% of patients on ECMO survived, with similar survival rates between VV and VA ECMO. Additional demographic characteristics and treatment parameters are outlined in Table 1.
Table 1.
Demographic and clinical characteristics of 67 adult patients on ECMO.
| Variable | Total (n=67) | No Severe Thrombocytopenia (n=41) | Severe Thrombocytopenia (n=26) | P-Value |
|---|---|---|---|---|
| Patient Demographics | ||||
| Age, years | 51.1 ± 15.9 | 51.6 ± 16.0 | 50.4 ± 16.0 | 0.759 |
| Male,n(%) | 44 (65.7) | 28 (68.3) | 16 (61.5) | 0.606 |
| Body Mass Index, kg/m2 | 32.7 ± 9.7 | 33.0 ± 10.3 | 32.3 ± 8.9 | 0.762 |
| Weight, kg | 99.2 ± 32.9 | 99.9 ± 34.4 | 98.2 ± 31.1 | 0.833 |
| Patients on Antiplatelet Drugs,n(%) | 17 (25.4) | 12 (29.3) | 5 (19.2) | 0.403 |
| Aspirin,n (%) | 16 (23.9) | 11 (26.8) | 5 (19.2) | 0.565 |
| Clopidogrel, n (%) | 7 (10.4) | 2 (4.9) | 5 (19.2) | 0.099 |
| Prasugrel, n (%) | 1 (1.5) | 1 (2.4) | 0 (0.0) | N/A |
| Mean Initial Platelet Count, x 109/L | 164.6 ± 91.5 | 199.0 ± 91.2 | 110.5 ± 61.9 | < 0.001† |
| Mean Platelet Count, x 109/L | 118.1 ± 49.6 | 122.0 ± 43.7 | 111.4 ± 58.1 | 0.414 |
| Mean aPTT, seconds | 95.3 ± 37.2 | 93.4 ± 38.6 | 98.4 ± 35.3 | 0.587 |
| Therapeutic anti-FXa*, % | 40.7 ± 25.0 | 40.3 ± 26.3 | 41.5 ± 23.1 | 0.854 |
| Mean Albumin, g/dL | 2.0 ± 0.5 | 2.1 ± 0.5 | 2.0 ± 0.5 | 0.421 |
| Reason for ECMO | ||||
| Post-cardiotomy, n (%) | 16 (23.9) | 5 (12.2) | 11 (42.3) | 0.008 † |
| Acute myocardial infarction, n (%) | 12 (17.9) | 10 (24.4) | 2 (7.7) | 0.109 |
| Decompensated heart failure, n (%) | 9 (13.4) | 6 (14.6) | 3 (11.5) | 1.000 |
| Sepsis, n (%) | 2 (3.0) | 1 (2.4) | 1 (3.8) | 1.000 |
| Refractory arrhythmia, n (%) | 2 (3.0) | 2 (4.9) | 0 (0.0) | N/A |
| Pulmonary embolism, n (%) | 2 (3.0) | 0 (0.0) | 2 (7.7) | N/A |
| Acute respiratory distress, n (%) | 17 (25.4) | 14 (34.1) | 3 (11.5) | 0.047 † |
| Pneumonia, n (%) | 4 (6.0) | 2 (4.9) | 2 (7.7) | 0.638 |
| Trauma, n (%) | 2 (3.0) | 1 (2.4) | 1 (3.8) | 1.000 |
| Parameters of ECMO Treatment | ||||
| Venoarterial ECMO, n (%) | 44 (65.7) | 24 (58.5) | 20 (76.9) | 0.187 |
| Average Flow Rate, mL/hour | 3846 ± 643 | 3889 ± 697 | 3776 ± 551 | 0.467 |
| Duration, days | 7.0 ± 5.0 | 7.4 ± 5.5 | 6.2 ± 4.2 | 0.310 |
| Mean Daily Heparin, units/day | 25427 ± 15341 | 25507 ± 15562 | 25301 ± 15290 | 0.958 |
| Average Number of Platelet Checks/Day | 3.8 ± 1.7 | 3.5 ± 1.4 | 4.1 ± 1.9 | 0.174 |
| Efficacy and Safety | ||||
| Incidence of Bleeding,n(%) | 45 (67.2) | 25 (61.0) | 20 (76.9) | 0.196 |
| ISTH Major Bleed,n(%) | 37 (55.2) | 19 (46.3) | 18 (69.2) | 0.082 |
| Thrombosis,n(%) | 16 (23.9) | 6 (14.6) | 10 (38.5) | 0.039 † |
| Number of RBCs | 8.9 ± 10.1 | 8.1 ± 9.7 | 10.3 ± 10.8 | 0.390 |
| Survive ECMO,n(%) | 39 (58.2) | 24 (58.5) | 15 (57.7) | 1.000 |
Significant at P=0.05
Data represents mean ± SD
Note: several patients received multiple antiplatelet agents while on ECMO
Within the cohort, patients on VA ECMO were older than those on VV ECMO (55.5 years vs. 42.7 years, P=0.002), with higher mean albumin levels (2.2 g/dL vs. 1.8 g/dL, P=0.016) and higher incidences of bleeding (77.3% vs. 47.8%, P=0.027). Patients on VV ECMO had higher mean platelet counts (138.8 × 109/L vs. 107.3 × 109/L, P=0.033), therapeutic anti-FXa levels (54.6% vs. 33.9%, P=0.002), days on ECMO (9.0 days vs. 5.9 days, P=0.033), and mean daily heparin use (35,268 units/day vs. 20,282 units/day, P=0.033). The average platelet count for the entire cohort at the time of first thrombotic event was 105 ± 45 × 109/L, and the average platelet count for individuals who had at least one instance of severe thrombocytopenia and developed a thrombus was 85 ± 32 × 109/L at the time of first thrombosis. Of the 69 patients in the cohort, 13 patients underwent testing for anti-platelet factor 4 (PF4)/heparin antibodies given suspicion for heparin induced thrombocytopenia (HIT). Four patients had a subsequent serotonin release assay (SRA) sent. Among these patients, 3 (21%) were PF4-positive and 1 (8%) had a positive confirmatory SRA. Patients undergoing cardiac surgery were more likely to have severe thrombocytopenia (P=0.008) while those with ARDS were less likely (P=0.047) (Table 1). The most common indications for ECMO were acute respiratory distress syndrome (25.4%), post-cardiotomy circulatory support (23.9%), acute myocardial infarction (19.4%), and decompensated heart failure (13.4%). Less common indications are also detailed in Table 1.
Thrombotic and Hemorrhagic Events
Thrombotic and hemorrhagic events were common within our cohort. Documented bleeding was recorded in 45 (67.2%) patients, and of these, 37 (55.2%) patients suffered a major bleed, as defined by the ISTH consensus criteria.7 Clinically evident thrombosis occurred in 16 (23.9%) patients. Additional clinical outcomes are described in Table 1. The mean number of units of red blood cells transfused was 8.9 units and 17 (25.4% of patients) received concurrent antiplatelet therapy while on ECMO. Due to the limitations of the retrospective analysis we were unable to evaluate the platelet transfusions.
Of our entire study population, thrombotic events most commonly occurred in arterial sites (10.4%), followed by venous (9.0%) Major hemorrhagic events were common in the cannula site (7.5%) and the GI tract (7.5%). Additional sites of thrombotic and hemorrhagic events are outlined in Table 2.
Table 2.
Site of thrombotic and hemorrhagic events.
| Variable | Total (n=67) | VA (n=44) | VV (n=23) | P-Value |
|---|---|---|---|---|
| Site of thrombotic event | ||||
| Venous | 6 (9.0) | 0 (0.0) | 6 (26.1) | N/A |
| Arterial | 7 (10.4) | 6 (13.6) | 1 (4.3) | 0.408 |
| Device | 1 (1.5) | 1 (2.3) | 0 (0.0) | N/A |
| Intracardiac | 3 (4.5) | 3 (6.8) | 0 (0.0) | N/A |
| Cerebrovascular accident | 1 (1.5) | 1 (2.3) | 0 (0.0) | N/A |
| Other | 1 (1.5) | 0 (0.0) | 1 (4.3) | N/A |
| Site of any hemorrhagic event | ||||
| Retroperitoneum | 2 (3.0) | 1 (2.3) | 1 (4.3) | 1.000 |
| Mediastinum | 2 (3.0) | 2 (4.5) | 0 (0.0) | N/A |
| GI tract | 6 (9.0) | 6 (13.6) | 0 (0.0) | N/A |
| Cannula site | 6 (9.0) | 5 (11.4) | 1 (4.3) | 0.656 |
| Other | 18 (26.9) | 12 (27.3) | 6 (26.1) | 1.000 |
| Pulmonary | 5 (7.5) | 2 (4.5) | 3 (13.0) | 0.330 |
| Surgical Site | 6 (9.0) | 6 (13.6) | 0 (0.0) | N/A |
| Site of major hemorrhagic event | ||||
| Retroperitoneum | 2 (3.0) | 1 (2.3) | 1 (4.3) | 1.000 |
| Mediastinum | 1 (1.5) | 1 (2.3) | 0 (0.0) | N/A |
| GI tract | 5 (7.5) | 5 (11.4) | 0 (0.0) | N/A |
| Cannula site | 5 (7.5) | 4 (9.1) | 1 (4.3) | 0.653 |
| Other | 17 (25.4) | 9 (20.5) | 8 (34.8) | 0.243 |
| Pulmonary | 3 (4.5) | 2 (4.5) | 1 (4.3) | 1.000 |
| Surgical Site | 3 (4.5) | 3 (6.8) | 0 (0.0) | N/A |
Note: several patients had multiple sites of thrombotic and hemorrhagic events while on ECMO
Incidence and Clinical Ramifications of Thrombocytopenia
In our cohort, 26 patients (38.8%) developed severe thrombocytopenia (platelet count <50 × 109/L) for at least one day while on ECMO. Of the 26 patients that developed severe thrombocytopenia, 20 (76.9%) were receiving VA ECMO and 6 (23.1%) were receiving VV ECMO. The mean daily heparin dose was not significantly different in patients who developed severe thrombocytopenia versus those who did not (25,301 vs. 25,507 units, P=0.958). Antiplatelet drug usage was not statistically significant between those who developed severe thrombocytopenia compared to those who did not (23.1% vs. 43.9%, P=0.403). Univariate analysis showed that developing severe thrombocytopenia significantly increased the odds of a thrombotic event (OR 3.65, 95% CI 1.13–11.77, P=0.031) and decreased the odds of surviving hospitalization (OR 0.288, 95% CI 0.100–0.835, P=0.022) (Table 3). Additionally, the development of severe thrombocytopenia significantly increased the odds of a major hemorrhagic event in patients on VA ECMO (OR 5.59, 95% CI 1.27–24.58, P=0.023). Using multivariate logistic regression and controlling for age, concurrent use of antiplatelet drugs, duration of ECMO, body mass index (BMI), and mean albumin level, we found that age and severe thrombocytopenia were risk factors for thrombotic events (Table 4).
Table 3.
Univariate logistic regression assessing the relationship between platelet count and relevant clinical outcomes.
| Outcome | Total (n=67) | VA (n=44) | VV (n=23) | ||||||
|---|---|---|---|---|---|---|---|---|---|
|
| |||||||||
| OR | 95% CI | P-Value | OR | 95% CI | P-Value | OR | 95% CI | P-Value | |
| Mean Platelet Count | |||||||||
| Thrombosis | 1.003 | 0.992–1.014 | 0.570 | 0.996 | 0.976–1.016 | 0.675 | 1.004 | 0.989–1.019 | 0.635 |
| Major Bleed | 1.001 | 0.991–1.011 | 0.807 | 0.995 | 0.981–1.010 | 0.541 | 1.010 | 0.995–1.027 | 0.198 |
| Incidence of Bleeding | 1.002 | 0.991–1.012 | 0.751 | 1.001 | 0.984–1.019 | 0.882 | 1.010 | 0.995–1.027 | 0.198 |
| Platelet Count Range | |||||||||
| Thrombosis | 0.992 | 0.983–1.001 | 0.098 | 0.995 | 0.982–1.008 | 0.422 | 0.991 | 0.978–1.004 | 0.166 |
| Major Bleed | 0.999 | 0.992–1.005 | 0.640 | 0.991 | 0.981–1.000 | 0.061 | 1.007 | 0.997–1.018 | 0.174 |
| Incidence of Bleeding | 1.001 | 0.994–1.007 | 0.869 | 0.993 | 0.984–1.003 | 0.170 | 1.007 | 0.997–1.018 | 0.174 |
| Severe Thrombocytopenia | |||||||||
| Thrombosis | 3.646 | 1.129–11.774 | 0.031 † | 4.714 | 0.832–26.723 | 0.080 | 6.500 | 0.850–49.687 | 0.071 |
| Major Bleed | 2.605 | 0.926–7.331 | 0.070 | 5.600 | 1.432–21.894 | 0.013 † | 0.444 | 0.063–3.112 | 0.414 |
| Incidence of Bleeding | 2.133 | 0.705–6.456 | 0.180 | 4.500 | 0.831–24.373 | 0.081 | 0.444 | 0.063–3.112 | 0.414 |
Significant at P=0.05
Table 4.
Multivariate logistic regression to predict outcome variables.
| Outcome | Total (n=67) | VA (n=44) | VV (n=23) | ||||||
|---|---|---|---|---|---|---|---|---|---|
|
| |||||||||
| OR | 95% CI | P-Value | OR | 95% CI | P-Value | OR | 95% CI | P-Value | |
| Predictors of Any Bleeding * | |||||||||
| Age | 1.02 | 0.98–1.06 | 0.400 | 1.06 | 0.98–1.14 | 0.170 | 0.91 | 0.83–1.01 | 0.070 |
| Antiplatelet Drugs | 0.60 | 0.15–2.34 | 0.459 | 0.47 | 0.07–2.89 | 0.411 | 0.44 | 0.02–9.65 | 0.602 |
| Duration of ECMO | 1.07 | 0.94–1.23 | 0.322 | 1.16 | 0.86–1.58 | 0.330 | 1.36 | 0.97–1.92 | 0.078 |
| Body Mass Index | 1.02 | 0.96–1.09 | 0.442 | 1.08 | 0.95–1.21 | 0.234 | 0.92 | 0.79–1.08 | 0.326 |
| Severe Thrombocytopenia | 2.46 | 0.74–8.22 | 0.144 | 6.61 | 0.93–46.84 | 0.059 | 0.25 | 0.01–5.24 | 0.375 |
| Albumin | 3.00 | 0.89–10.16 | 0.077 | 5.43 | 0.86–8.08 | 0.073 | 0.08 | 0.00–4.82 | 0.223 |
| Predictors of Major Bleeds * | |||||||||
| Age | 0.99 | 0.96–1.03 | 0.65 | 1.00 | 0.94–1.07 | 0.904 | 0.91 | 0.83–1.01 | 0.070 |
| Antiplatelet Drugs | 0.35 | 0.09–1.35 | 0.13 | 0.22 | 0.04–1.21 | 0.082 | 0.44 | 0.02–9.65 | 0.602 |
| Duration of ECMO | 1.05 | 0.93–1.18 | 0.40 | 0.95 | 0.76–1.18 | 0.632 | 1.36 | 0.97–1.92 | 0.078 |
| Body Mass Index | 1.00 | 0.95–1.06 | 0.93 | 1.00 | 0.94–1.08 | 0.889 | 0.92 | 0.79–1.08 | 0.326 |
| Severe Thrombocytopenia | 2.68 | 0.87–8.25 | 0.09 | 5.59 | 1.27–24.58 | 0.023 † | 0.25 | 0.01–5.24 | 0.375 |
| Albumin | 1.67 | 0.55–5.04 | 0.37 | 1.77 | 0.39–8.08 | 0.463 | 0.08 | 0.00–4.82 | 0.223 |
| Predictors of Thrombosis * | |||||||||
| Age | 0.94 | 0.90–0.99 | 0.011 † | 0.95 | 0.87–1.03 | 0.216 | 0.91 | 0.80–1.04 | 0.165 |
| Antiplatelet Drugs | 2.49 | 0.49–12.70 | 0.274 | 4.50 | 0.48–42.09 | 0.187 | 1.39 | 0.05–37.61 | 0.844 |
| Duration of ECMO | 1.03 | 0.91–1.18 | 0.629 | 0.93 | 0.69–1.24 | 0.608 | 0.98 | 0.79–1.21 | 0.847 |
| Body Mass Index | 1.00 | 0.93–1.08 | 0.945 | 0.93 | 0.82–1.06 | 0.258 | 1.08 | 0.90–1.30 | 0.401 |
| Severe Thrombocytopenia | 5.30 | 1.30–21.68 | 0.02 † | 5.82 | 0.74–46.09 | 0.095 | 665.09 | 0.71–620026.09 | 0.062 |
| Albumin | 1.20 | 0.35–4.09 | 0.771 | 0.78 | 0.12–5.17 | 0.799 | 146.33 | 1.16–18499.85 | 0.043 † |
| Predictors of Severe Thrombocytopenia ** | |||||||||
| Age | 1.00 | 0.96–1.03 | 0.933 | 0.98 | 0.93–1.04 | 0.561 | 0.94 | 0.85–1.04 | 0.234 |
| Antiplatelet Drugs | 0.50 | 0.13–1.94 | 0.316 | 0.58 | 0.13–2.57 | 0.475 | 0.00 | 0.00-Inf | 0.996 |
| Duration of ECMO | 0.91 | 0.80–1.04 | 0.182 | 0.98 | 0.80–1.21 | 0.879 | 0.76 | 0.49–1.18 | 0.220 |
| Body Mass Index | 0.98 | 0.93–1.04 | 0.521 | 0.98 | 0.92–1.05 | 0.627 | 1.12 | 0.94–1.33 | 0.214 |
| Albumin | 0.77 | 0.26–2.25 | 0.628 | 0.71 | 0.17–2.89 | 0.632 | 0.00 | 0.00–1.22 | 0.058 |
ECMO, extracorporeal membrane oxygenation; Body Mass Index, body mass index; OR, odds ratio; CI, confidence interval. For continuous variables, odds ratios are in terms of changes in odds as a result of a one-unit change in the variable.
Significant at P=0.05
Adjusted for age, antiplatelet drugs, duration of ECMO, Body Mass Index, severe thrombocytopenia, and albumin level
Adjusted for age, antiplatelet drugs, duration of ECMO, Body Mass Index, and albumin level
The mean platelet count lay outside the physiologic range (150–400 × 109/L) nearly 100% of the time and tended to decrease over time on ECMO (Figure 1). When analyzed by type of ECMO, patients on VA ECMO trended consistently below the physiologic range, whereas patients on VV experienced an initial dip followed by a rise in platelet count. No statistically significant difference in platelet count was observed between patients with and without thrombotic events (P=0.630) (Figure 1).
Figure 1.
Platelet counts stratified by VA and VV ECMO. Consecutive platelet counts while undergoing ECMO treatment (A-C). Platelet counts stratified by hemorrhagic (D-F) and thrombotic (G-I) events. The gray region represents the physiologic platelet count range (150–400 × 109/L).
Discussion
In the outlined cohort study of 67 patients, we found that severe thrombocytopenia was common in patients undergoing ECMO and is predictive of mortality. After controlling for possible confounding variables in multivariate analysis, severe thrombocytopenia also predicted major bleeding in VA ECMO (although not in the entire cohort), and surprisingly was a risk factor for thrombotic events in the entire cohort. Our study provides new insight on the prevalence and clinical consequences of severe thrombocytopenia in patients on ECMO that have not previously been reported. Of the patients in our study, 38.8% developed severe thrombocytopenia. Notably, of the 26 patients that developed severe thrombocytopenia, 20 (76.9%) were receiving VA ECMO and 6 (23.1%) were receiving VV ECMO. While limited by the small sample size, the duration of ECMO was not associated with the development of thrombocytopenia, consistent with previous studies.5,8
Thrombocytopenia is not uncommon in critically ill patients. Prior analysis have suggested incidences of thrombocytopenia up to 50% at some time point in critically ill patients admitted to the intensive care unit, with 5% to 20% of patients developing severe thrombocytopenia.9 Multiple potentially overlapping mechanisms have been suggested to explain the worsening thrombocytopenia seen in critically ill patients including platelet clumping, hemodilution, underproduction, destruction or consumption.10 Prior analysis have also suggests thrombocytopenia to be a predictor of both mortality and hemorrhage in critically ill patients, similar to our findings in patients on ECMO.11 While disproportionately composed of patients on VA-ECMO, the overall survival of 58% in our cohort is similar to the survival reported by the Extracorporeal Life Support Organization Registry International Report 2016 of 58.7% in all adults on VV and VA ECMO.12
At first glance, it may seem counter intuitive that subjects with severe thrombocytopenia developed more thrombotic complications; however, there is some literature to suggest that the association between thrombocytopenia and ECMO use is confounded by the subjects’ severity and duration of critical illness. One study found that 22% of subjects developed severe thrombocytopenia while receiving ECMO, but they also showed that a higher initial severity of critical illness, lower baseline platelet count, and the development of hepatic or renal failure accounted for that association. It is possible that severe thrombocytopenia develops in patients who are more systemically ill and in more severe underlying proinflammatory and prothrombotic states.8Alternatively, it is plausible that the finding of severe thrombocytopenia led to less aggressive anticoagulation strategies and increased risk of thrombosis. Indeed, while the level of anticoagulation appeared to be unaffected in our cohort, the use of antiplatelet therapy was numerically diminished in patients who developed severe thrombocytopenia. Lastly, as we did not evaluate the temporal relationship of thrombocytopenia to thrombosis, thrombocytopenia may be a result of thrombosis rather than a cause. Equally plausible is the possibility that thrombocytopenia was a harbinger of still-occult thrombosis.
In regards to bleeding, several studies have demonstrated impaired platelet function during ECMO. Reduced expression of adhesion receptors GPIbα and GPVI and decreased levels of biomarkers of platelet activation, PF4 and ß-thromboglobulin during ECMO have been suggested as mechanisms for disrupted functional and structural thrombocyte integrity and aggregation in ECMO patients.15,16
The findings from this study are hypothesis generating and clinically relevant for several reasons. First, thrombosis and hemorrhage are common and severe complications of ECMO, and while acquired platelet dysfunction has previously been implicated,17 platelet count remains a practical variable that can be easily tracked, and adjusted with transfusion or risk factor modification. However, previous studies have suggested that platelet transfusion may paradoxically increase the risk of thrombosis.18 For instance, Cashen et al. used a multivariable analysis to demonstrate that platelet transfusion volume was independently associated with increased daily thrombotic risk.19
Additionally, the role of anticoagulation in ECMO is not fully clear, given increasing reports of ECMO ran without anticoagulation highlighting the complex hemostatic milieu brought about by blood-device interactions and non-physiologic blood rheology provoked by the extracorporeal circuit.23,24 As anticoagulant free ECMO seems increasingly feasible, it may seem plausible that severe thrombocytopenia is an indication to safely pause anticoagulation. However, our study found that the differences in mean daily heparin dose and the use of antiplatelet drugs were not statistically significant in patients who developed severe thrombocytopenia versus those who did not (P=0.958, P=0.403), suggesting that, at least at our center, there is a perceived need to continue heparin as much as possible. Although the ideal anticoagulation strategy for patients on ECMO therapy remains unclear, recommendations from the Extracorporeal Life Support Organization (ELSO) recommend UFH in all patients on ECMO.25–27 A clinical trial is underway comparing traditional anticoagulation to anticoagulant free VV-ECMO which may provide insight into the true risks and benefits of anticoagulation in this population (NCT04273607). While further analysis is needed before definitive conclusions can be drawn, the finding that severe thrombocytopenia is associated with thrombosis suggests that continuing anticoagulation may be a consideration in ECMO patients who develop severe thrombocytopenia.
These findings should be taken in the context of the limitations of the analysis, including the small size of our cohort and the retrospective nature of our analysis. The generalizability of these findings is likely constrained due to the small sample size. Furthermore, the data were collected retrospectively and therefore, one cannot control for all clinical factors that may have affected outcomes or may serve as residual confounders. A proportion of ECMO patients may have received left ventricular impellas as well, which were unaccounted for in this analysis. As impellas may represent an independent risk factor for bleeding, thrombosis, and thrombocytopenia this may be a confounder to our analysis. Moreover, due to the limitations of our data capture methods, we considered only the first 30 consecutive platelet counts when evaluating if patients developed severe thrombocytopenia; however, only two patients developed severe thrombocytopenia for a non-sustained period of time beyond the first 30 counts. In addition, asymptomatic ECMO patients were not routinely screened for thrombotic events; it is possible that more event would have been observed with this diagnostic strategy. Future studies should also incorporate validated ECMO risk scores such as APACHE II, SAVE, or PRESERVE to provide further information on disease severity and prediction of survival. Nonetheless, the results from this analysis should be considered as a larger prospective analysis evaluating the complex derangements of hemostasis and thrombosis associated with ECMO. As further data are generated, clinicians should be conscious of platelet count in patients on ECMO and consider potential strategies to mitigate the significant risks of hemorrhage and thrombosis associated with ECMO.
Funding
This work was supported by the National Heart, Lung, and Blood Institute under Grants HL101972, HL144113, and HL151367.
Footnotes
Declaration of Interests
J. J. Shatzel is a consultant for Aronora, Inc. The remaining authors report no conflict of interest.
References
- 1.Aubron C, DePuydt J, Belon F, et al. Predictive factors of bleeding events in adults undergoing extracorporeal membrane oxygenation. Ann Intensive Care. 2016;6(1):97–97. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Raffini L. Anticoagulation with VADs and ECMO: walking the tightrope. Hematology Am Soc Hematol Educ Program. Dec 8 2017;2017(1):674–680. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Kato C, Oakes M, Kim M, et al. Anticoagulation Strategies in Extracorporeal Circulatory Devices in Adult Populations. European Journal of Haematology. 2020. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Menaker J, Tabatabai A, Rector R, et al. Incidence of Cannula-Associated Deep Vein Thrombosis After Veno-Venous Extracorporeal Membrane Oxygenation. ASAIO J. Sep/Oct 2017;63(5):588–591. [DOI] [PubMed] [Google Scholar]
- 5.Jiritano F, Serraino GF, Ten Cate H, et al. Platelets and extra-corporeal membrane oxygenation in adult patients: a systematic review and meta-analysis. Intensive Care Med. Jun 2020;46(6):1154–1169. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Schulman S, Kearon C. Definition of major bleeding in clinical investigations of antihemostatic medicinal products in non-surgical patients. J Thromb Haemost. Apr 2005;3(4):692–694. [DOI] [PubMed] [Google Scholar]
- 7.Schulman S, Kearon C, Subcommittee on Control of Anticoagulation of the S, Standardization Committee of the International Society on T, Haemostasis. Definition of major bleeding in clinical investigations of antihemostatic medicinal products in non-surgical patients. J Thromb Haemost. Apr 2005;3(4):692–694. [DOI] [PubMed] [Google Scholar]
- 8.Abrams D, Baldwin MR, Champion M, et al. Thrombocytopenia and extracorporeal membrane oxygenation in adults with acute respiratory failure: a cohort study. Intensive Care Med. May 2016;42(5):844–852. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Greinacher A, Selleng S. How I evaluate and treat thrombocytopenia in the intensive care unit patient. Blood. Dec 29 2016;128(26):3032–3042. [DOI] [PubMed] [Google Scholar]
- 10.Thiele T, Selleng K, Selleng S, Greinacher A, Bakchoul T. Thrombocytopenia in the intensive care unit-diagnostic approach and management. Semin Hematol. Jul 2013;50(3):239–250. [DOI] [PubMed] [Google Scholar]
- 11.Zarychanski R, Houston DS. Assessing thrombocytopenia in the intensive care unit: the past, present, and future. Hematology Am Soc Hematol Educ Program. Dec 8 2017;2017(1):660–666. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Thiagarajan RR, Barbaro RP, Rycus PT, et al. Extracorporeal Life Support Organization Registry International Report 2016. ASAIO Journal. 2017;63(1):60–67. [DOI] [PubMed] [Google Scholar]
- 13.Boddi M, Peris A. Deep Vein Thrombosis in Intensive Care. Adv Exp Med Biol. 2017;906:167–181. [DOI] [PubMed] [Google Scholar]
- 14.Hui P, Cook DJ, Lim W, Fraser GA, Arnold DM. The frequency and clinical significance of thrombocytopenia complicating critical illness: a systematic review. Chest. Feb 2011;139(2):271–278. [DOI] [PubMed] [Google Scholar]
- 15.Lukito P, Wong A, Jing J, et al. Mechanical circulatory support is associated with loss of platelet receptors glycoprotein Ibα and glycoprotein VI. J Thromb Haemost. Nov 2016;14(11):2253–2260. [DOI] [PubMed] [Google Scholar]
- 16.Strauss R, Wehler M, Mehler K, Kreutzer D, Koebnick C, Hahn EG. Thrombocytopenia in patients in the medical intensive care unit: bleeding prevalence, transfusion requirements, and outcome. Crit Care Med. Aug 2002;30(8):1765–1771. [DOI] [PubMed] [Google Scholar]
- 17.Kalbhenn J, Schlagenhauf A, Rosenfelder S, Schmutz A, Zieger B. Acquired von Willebrand syndrome and impaired platelet function during venovenous extracorporeal membrane oxygenation: Rapid onset and fast recovery. J Heart Lung Transplant. Aug 2018;37(8):985–991. [DOI] [PubMed] [Google Scholar]
- 18.Baharoglu MI, Cordonnier C, Al-Shahi Salman R, et al. Platelet transfusion versus standard care after acute stroke due to spontaneous cerebral haemorrhage associated with antiplatelet therapy (PATCH): a randomised, open-label, phase 3 trial. Lancet. Jun 25 2016;387(10038):2605–2613. [DOI] [PubMed] [Google Scholar]
- 19.Cashen K, Dalton H, Reeder RW, et al. Platelet Transfusion Practice and Related Outcomes in Pediatric Extracorporeal Membrane Oxygenation. Pediatr Crit Care Med. Feb 2020;21(2):178–185. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Sylman JL, Mitrugno A, Tormoen GW, Wagner TH, Mallick P, McCarty OJT. Platelet count as a predictor of metastasis and venous thromboembolism in patients with cancer. Converg Sci Phys Oncol. Jun 2017;3(2). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Sylman JL, Boyce HB, Mitrugno A, et al. A Temporal Examination of Platelet Counts as a Predictor of Prognosis in Lung, Prostate, and Colon Cancer Patients. Sci Rep. Apr 26 2018;8(1):6564. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Sylman JL, Mitrugno A, Atallah M, et al. The Predictive Value of Inflammation-Related Peripheral Blood Measurements in Cancer Staging and Prognosis. Front Oncol. 2018;8:78. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Kurihara C, Walter JM, Karim A, et al. Feasibility of Venovenous Extracorporeal Membrane Oxygenation Without Systemic Anticoagulation. Ann Thorac Surg. Mar 12 2020. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Olson SR, Murphree CR, Zonies D, et al. Thrombosis and Bleeding in Extracorporeal Membrane Oxygenation (ECMO) Without Anticoagulation: A Systematic Review. ASAIO J. Jul 23 2020. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Combes A, Brodie D, Chen YS, et al. The ICM research agenda on extracorporeal life support. Intensive Care Med. Sep 2017;43(9):1306–1318. [DOI] [PubMed] [Google Scholar]
- 26.Sy E, Sklar MC, Lequier L, Fan E, Kanji HD. Anticoagulation practices and the prevalence of major bleeding, thromboembolic events, and mortality in venoarterial extracorporeal membrane oxygenation: A systematic review and meta-analysis. J Crit Care. Jun 2017;39:87–96. [DOI] [PubMed] [Google Scholar]
- 27.Organization ELS. Extracorporeal Life Support Organization (ELSO) General Guidelines for All ECLS Cases. 2017; Version 1.4:https://www.elso.org/Portals/0/ELSO%20Guidelines%20General%20All%20ECLS%20Version%201_4.pdf, 2020.