Greater Fibrinolysis Resistance but No Greater Platelet Aggregation in Critically Ill COVID-19 Patients

Patients diagnosed with COVID-19 and admitted to the intensive care unit (ICU) of the authors’ institution were included into this prospective, monocentric observational study. The inclusion criteria were age greater than or equal to 18 yr, moderate to severe acute respiratory distress syndrome (ARDS) due to SARS-CoV-2 infection, and no history of congenital, acquired, or any other known coagulopathy.

The study was performed in accordance with the Declaration of Helsinki. Approval from the local ethics committee was obtained before the study was conducted (No. 20-643), and a waiver regarding the requirement of written informed consent from COVID-19 patients was authorized. All participants of the control group provided written informed consent and were recruited only for the current study. Patient care and study conduct complied with good clinical practice.

Demographic and biochemical data as well as the medical history of patients admitted for COVID-19 were recorded. A healthy control population comprising volunteers without previous history of hyper- or hypocoagulable disorders was examined for this study. These individuals were recruited concurrently with the patients from the community between April 1 and April 15, 2020.

Upon admission to the ICU, all patients received mechanical ventilation and critical care therapy as put forth by Poston et al.⁷  The regimen of thrombosis prophylaxis was 60 mg (or 80 mg at body mass index greater than 35 kg/m²) of low-molecular-weight heparin (calcium enoxaparin) twice a day. In the case of vasopressor therapy, the regimen was changed to administration of unfractionated heparin with a target activated partial thromboplastin time (PTT) of 50 to 70 s. Antithrombin (AT) concentrate was replaced regularly to maintain a level of 80% or greater. No additional drugs with known antiplatelet effects were taken other than indicated. No experimentally intended antiviral therapies (remdesivir, hydroxychloroquine, or other antiviral agents) were applied.

Venous blood was collected via a cannula inserted into a cubital vein. Collection tubes for conventional coagulation analysis were prefilled with sodium citrate (S-Monovette 1.8 ml, sodium-citrate 3.2% [1:10]; Sarstedt AG, Germany) and analyzed by an ACL Top 700 CTS (Werfen GmbH, Spain). Hematological analyses were performed using collection tubes prefilled with ethylenediaminetetraacetate (S-Monovette 1.6 ml, K3 EDTA; Sarstedt AG) and analyzed by an XN 9000 (Sysmex GmbH, Germany). Platelet count was determined by fluorescence flow cytometry on an XE 2100 (Sysmex GmbH), and biochemical parameters were assessed using serum collection tubes (S-Monovette 7.5 ml, Serum Gel with clotting activator; Sarstedt AG) and analyzed by a Cobas 8000 (Roche Diagnostics, Germany).

For ClotPro analysis, blood was collected into collection tubes prefilled with sodium citrate (S-Monovette 1.8 ml, citrate 3.2% [1:10]; Sarstedt AG). For multiple electrode aggregometry analysis, a heparinized blood gas analysis sample tube (safePICO; Radiometer, Germany) was used.

Platelet function was measured by multiple electrode aggregometry using the Multiplate analyzer 15 min after blood draw and after activation with commercially available standard reagents (Roche, Basel, Switzerland) as previously published.⁸  Blood samples were analyzed at 37°C. To test different methods of aggregation induction, aggregation was stimulated via (1) adenosine diphosphate ([ADP] 6.4 mmol/l) receptors by ADP; (2) arachidonic acid, the substrate of cyclooxygenase (0.5 mmol/l arachidonic acid), which subsequently forms the potent platelet activator thromboxane A₂; and (3) thrombin receptor activating peptide 6 (32 mmol/l) via the platelet surface platelet receptor as described previously.⁸  To identify abnormal values in multiple electrode aggregometry assays, reference ranges were defined in accordance with the manufacturer’s recommendations for heparinized blood samples.⁹ 

Thromboelastometric assays were performed 15 min after blood draw. Blood samples were analyzed at 37°C using ClotPro analyzer (enicor GmbH, Haemonetics, Germany).

The ClotPro analyzer provides bedside viscoelastometric measurements of whole blood coagulation by recording kinetic changes in a sample of citrated whole blood, similar to rotational thromboelastometry (ROTEM).^10,11  The blood sample is placed into a cylindrical cup, which rotates alternately. A stationary cylindrical pin is then inserted into the cup. The clotting sample reduces the movement of the cup gradually as the clot firmness rises. The cup movement is recorded and transformed into an amplitude, which is continuously recorded against the time and expressed in millimeters for historical reasons.

The run time is set to 40 min and automatically stopped by the software. Regular quality control tests were run in accordance with the manufacturer’s instructions. For the current study, four tests were carried out using reagents provided by the manufacturer: recombinant tissue factor–triggered extrinsic pathway, which evaluates the extrinsic pathway; ellagic acid–activated intrinsic pathway, which evaluates the intrinsic pathway; cytochalasin D and synthetic glycoprotein IIb/IIIa antagonist, which are inhibitors of the rearrangement of microtubules in platelets and thus of platelet aggregation, which evaluates the contribution of fibrin to clot firmness; and tissue factor–triggered extrinsic pathway and activation of fibrinolysis by high-dose (650 ng/ml) recombinant tissue plasminogen activator (Tpa), which reflects resistance to fibrinolysis. All tissue factor–triggered assays contain polybrene as an antagonist of heparin present in the sample.

When performing viscoelastic tests, the following parameters were calculated: clotting time, which is the time from the start of the test until the clotting of the sample (2-mm clot firmness), expressed in seconds; clot formation time, which is the time from clotting time until an amplitude of 20 mm is detected, expressed in seconds; maximum clot firmness expressed in millimeters; maximum lysis of the clot in percentage of the maximum clot firmness; and the lysis time, which is the time from the end of clotting time until a lysis of 50% of the maximum clot firmness is recorded, expressed in seconds. The maximum lysis reflects the percentage of lysis in relation to the maximum clot firmness. For technical reasons, the lowest amplitude during lysis is 1.5 mm (i.e., if an amplitude of less than 1.5 mm is reached, the amplitude is still displayed as 1.5 mm). In conclusion, maximum lysis cannot reach 100%. To identify abnormal values in thromboelastometric assays, reference ranges were defined in accordance with the manufacturer’s instructions. The reference range for lysis time as given by the manufacturer ranges from 145 to 438 s.

No statistical power calculation was conducted before the study. The sample size was based on the available data. The current study is a post hoc analysis. Data were tested for normality using the Kolmogorov–Smirnov test. Data comparisons of patient characteristics were made using the Student’s t test; results of both multiple electrode aggregometry and thromboelastometry were made using the Mann–Whitney U test and Hodges–Lehmann estimator. Adjusted analysis was performed for the three main potential confounders. Here a stratified nonparametric approach was used for sex and a linear regression for the continuous confounders age and body mass index.

All statistical tests were two-tailed, and results with P < 0.05 were considered statistically significant. All calculations/analyses were performed with SPSS (Version 25; IBM, USA), R software (Version 3.6; The R Foundation, USA), and GraphPad Prism (Version 8; GraphPad, USA). There were no missing data.

The patient population included 27 patients diagnosed with COVID-19 and ARDS who were treated in our ICU. The demographic data are presented in table 1. Overall, 21 of the patients were male. Of all patients, 25 (93%) were obese according to the definition of the World Health Organization, and six (22%) had a body mass index greater than 40 kg/m². The mean age was 60 ± 13 yr. The mean ICU stay from admission until impedance aggregometric and viscoelastic assessment was 7 ± 3.5 days. The most frequent preexisting comorbidities were arterial hypertension (52%) and diabetes (41%). All patients were mechanically ventilated, and 14 patients (52%) received renal replacement therapy due to acute renal failure. To classify the severity of disease, the Simplified Acute Physiology Score II was assessed, revealing a mean score of 42.¹⁰  The control group consisted of 12 healthy volunteers without any known preexisting conditions. The mean age of the control group was 38 ± 6 years; the mean body mass index was 24.5 ± 2.6 kg/m², and there were statistically significant differences in age (P < 0.0001), body mass index (P < 0.0001), prevalence of preexisting conditions, use of mechanical ventilation (P < 0.0001), and renal replacement therapy (P < 0.0001) between the COVID-19 patients and the control group.

The results of the coagulation parameters are presented in table 2. The median values revealed that PTT; AT and fibrinogen levels; platelet count; PTT; and activity of factor II, factor V, factor VII, factor XI, factor XII, and protein C were within the normal range in patients with COVID-19. In 12 patients (44%), AT was substituted as mentioned above. The mean cumulative dose was 2,500 IU (data not shown). Furthermore, we detected elevated D-dimer levels (3,656 ng/ml [interquartile range, 1,130 to 6,749]), elevated activity of factor VIII (261.8 ± 78.7%) factor IX (150.7 ± 53.3%), and von Willebrand factor (vWF) antigen (554 ± 109.7%), and lesser activity of factor XIII (64.5 ± 35.1%) and protein S (45 ± 30.1%). Thrombocytosis (defined as platelet count greater than 350 10³/µl) was detected in 10 (37%) patients. Median D-dimer levels and activity of factor VIII, factor IX, and vWF antigen exceeded the upper reference limit. The median activity of factor X, factor XIII, and protein S was less than the lower reference limit.

Furthermore, we detected elevated median values of serum levels for c-reactive protein (16.7 ± 10.75 mg/dl), procalcitonin (1.7 ± 13.18 ng/ml), interleukin-6 (168 ± 904 pg/ml), ferritin (1,235 ± 423 ng/ml), and lactate dehydrogenase (494 ± 173 U/l), which are displayed in table 2.

For impedance aggregometric assays, eight patients on therapy with acetylsalicylic acid were excluded for the arachidonic acid test, and one patient with thrombopenia was excluded from the analysis in accordance with Hanke et al.¹¹  demonstrating multiple electrode aggregometry results being dependent on platelet count.

In patients with COVID-19, impedance aggregometric assays were performed for AUC of ADP (68 ± 37 U), AUC of arachidonic acid (102 ± 54 U), and AUC of thrombin receptor activating peptide (114 ± 61 U). In 12 patients, the results of AUC for ADP were less than the lower reference range.

Comparing the multiple electrode aggregometry results of COVID-19 patients and healthy controls demonstrated significantly lower results for mean AUC for ADP (68 ± 37 U vs. 91 ± 29 U; −31 [95% CI, −48 to −1]; P = 0.043) in COVID-19 patients. No significant differences between COVID-19 patients and healthy controls were observed for mean AUC for arachidonic acid and mean AUC for thrombin receptor–activating peptide (table 3, fig. 1 ). When checking for potential confounding effects of sex, age, and body mass index, the significance of the group difference for AUC for ADP may be explained by sex as it becomes insignificant after stratification.

Thromboelastometric analyses are presented in table 3 and figure 2. In detail, thromboelastometry revealed values below the lower reference range for clot formation time in extrinsic activation in 2 of 27 patients, for clot formation time in intrinsic activation in 5 of 27 patients and for maximum clot firmness in intrinsic activation in 1 of 27 patients. Thromboelastometry displayed values greater than the reference range for clot formation time in extrinsic activation in no patient, for clot formation time in intrinsic activation in 4 of 27 patients, and for maximum clot firmness in intrinsic activation in 14 of 27 patients.

Comparing the results from thromboelastometric assays of COVID-19 patients and healthy controls, significant differences were observed in extrinsic activation assay for the clotting time (mean ± SD, 88 ± 22 s vs. 60 ± 7 s; 22 [95% CI, 15 to 33]; P < 0.001) and the maximum clot firmness (68 ± 5 mm vs. 57 ± 4 mm; 11 [95% CI, 8 to 14]; P < 0.001).

Further, significant differences were observed in intrinsic activation assay for the clotting time (262 ± 120 s vs. 163 ± 12 s; 47 [95% CI, 25 to 92]; P < 0.001) and the maximum clot firmness (64 ± 8 mm vs. 56 ± 3 mm; 9 [95% CI, 5 to 13]; P = 0.001).

We also identified differences in the contribution of fibrin to clot firmness assay for the clotting time (104 ± 31 mm vs. 69 ± 14 mm; 28 [95% CI, 16 to 46]; P < 0.001) and the maximum clot firmness (37 ± 11 mm vs. 15 ± 4 mm; 21 [95% CI, 17 to 26]; P < 0.001).

Further, significant differences were found in the assay analyzing the extrinsic activation and activation of fibrinolysis by Tpa for the clotting time (68 ± 21 s vs. 42 ± 9 s; 25 [95% CI, 12 to 39]; P < 0.001), the maximum clot firmness (51 ± 12 mm vs. 26 ± 9 mm; 27 [95% CI, 19 to 33]; P < 0.001), the maximum lysis (93 ± 15% vs. 92 ± 4%; 3 [95% CI, 2 to 5]; P = 0.001), and the lysis time (530 ± 327 s vs. 211 ± 80 s; 238 [95% CI, 160 to 326]; P < 0.001).

Effect sizes (Cohen’s d) were calculated for the not significantly different tests and interpreted per Cohen¹²  as follows: AUC for arachidonic acid, −0.29/small effect; AUC for thrombin receptor activating peptide, −0.62/medium effect; clot formation time in extrinsic activation, −0.054/medium effect; and clot formation time in intrinsic activation, 0.44/medium effect.

When checking for associations with the potential confounders age and body mass index, these were not significant in any bivariable analyses. Stratified analysis with respect to sex in markers significantly associated with sex (maximum clot firmness in contribution of fibrin to clot firmness assay and maximum clot firmness in assay of extrinsic activation and activation of fibrinolysis by Tpa) did not change the significance of the results.

Our prospective, observational study of 27 patients with COVID-19 infection and moderate to severe ARDS revealed greater fibrinolysis resistance as reflected by thromboelastometry and no greater platelet aggregability using impedance aggregometric testing.

Assessment of coagulation factors and conventional coagulation parameters demonstrated elevated D-dimer levels as described previously and typical to COVID-19.²  Moreover, we identified a PTT within normal ranges; more vWF antigen, factor VIII, and factor IX; and less protein S, indicative of complement pathway activation, acute phase response, and an association with a procoagulant state. Considering the recent findings of endotheliitis¹³  in COVID-19 patients, the elevation of vWF levels may mirror endothelial activation or damage. Such elevated vWF levels, as recently described by others,^14,15  may therefore be considered as a surrogate parameter of endothelial dysfunction, supporting the procoagulant imbalance with a potentially higher risk of venous thromboembolism. The elevated factor IX levels cannot be conclusively explained by the data obtained. Since both individual variability¹⁶  and advanced age¹⁷  have been observed in association with higher levels of factor IX, no reliable differentiation can be drawn. Nevertheless, we would like to highlight the proven significance of factor VIII and factor IX elevations in regard to an associated high risk of venous thrombotic events,^18,19  which further highlights the need for a sophisticated anticoagulation regimen in these patients. It remains speculative whether the reported lower levels of factor XIII are acquired; such low levels may result from either an increased consumption or a reduced production and should be addressed in future studies.

In contrast with the findings of Ranucci et al.,²⁰  the fibrinogen values of the COVID-19 patients rarely exceeded but were close to the upper reference range, which contributed to the clinically significantly greater clot strength detected. While we observed a prolonged clotting time in intrinsic activation for patients suffering from COVID-19, this finding may in part have been altered by treatment with unfractionated heparin in eight patients. The observed higher levels of factor VIII and IX suggest that the prolonged clotting time in the intrinsic activation assay is not related to a factor deficiency. However, the lower factor XII level might have contributed to the prolonged clotting time of this test. Further, the results obtained within the assay of extrinsic activation and activation of fibrinolysis by Tpa should not be affected by unfractionated heparin or low-molecular-weight heparin, as it is activated via the extrinsic pathway (tissue factor) and contains a heparin antagonist (polybrene).

The analysis of impedance aggregometric assays revealed values within the normal reference ranges. In comparison to healthy controls, mean results of AUC for ADP were significantly lower, which is in line with changes in platelet aggregation in bacterial sepsis.²¹  In our patient cohort, we observed thrombocytosis in the majority of cases, in contrast to recently published data describing thrombocytopenia as a common finding in patients with severe COVID-19 infection.^22,23  Given that the analyses in this study were carried out on a single occasion, the most likely cause is reactive thrombocytosis, which has been reported for COVID-19. Impedance aggregometric measurements revealed no greater platelet aggregation, although it might have been suspected given the increased reports of thromboembolic events.^4,24  However, definitive conclusions on the platelet function in patients suffering from COVID-19 may not be drawn from the analysis presented within this manuscript, as sophisticated analyses including receiver operating characteristics curves and serial measurement during the hospital stay from a bigger patient population are warranted to further substantiate the current findings.

The thromboelastometric results of our study using the ClotPro reinforce and complement the current concept of a hypercoagulable pattern in COVID-19 patients with a profound derangement of hemostasis.²⁰  Hence, the performance of the recently introduced assay of extrinsic activation and activation of fibrinolysis by Tpa revealed new and relevant results.²⁵  Using the same reagents as in the extrinsic activation assay, this test is based on additional recombinant Tpa to induce fibrinolysis. Thus, the presence of lysis inhibitors (e.g., tranexamic acid) and their influence on blood clotting as well as Tpa-induced lysis can be assessed. Our results of the lysis time of this test indicate that there is a greater fibrinolysis resistance in COVID-19 patients in comparison to healthy controls.

In addition to the reported elevations in serum D-dimer levels for COVID-19 reflecting an activation of the coagulation system consistent with the described various clinical thrombotic events in these patients,^4,24,26  our results of the assay of extrinsic activation and activation of fibrinolysis by Tpa might therefore reflect a greater fibrinolysis resistance, which reinforces the procoagulant state described and complements the results of recently published viscoelastic measurements.^14,27  Of note, hypofibrinolysis or fibrinolysis shutdown is characterized by decreased fibrinolysis in assays without Tpa challenge.^21,28–30 

Moreover, the findings of our analyses are consistent with emerging observations suggesting that COVID-19 has features distinct from typical ARDS. In addition to the considerably well-preserved lung mechanics despite a severe hypoxemia, as characterized by high respiratory compliance, high shunt fraction, and prolonged mechanical ventilation,³¹  Magro et al. reported an association with the involvement of complement components within the pulmonary septal microvasculature, which is atypical for classic ARDS.³²  Such extensive complement involvement may lead to complex-mediated microvascular endothelial cell injury and subsequent activation of the coagulation pathway, which might explain our findings.³³  In addition, recently published findings of direct viral infection of endothelial cells and diffuse endothelial inflammation leading to endothelial dysfunction might further contribute to the observed procoagulant state of hemostasis.¹³ 

Dysregulated fibrinolysis, often observed in the context of critical illness,^34,35  can lead to a so-called “fibrinolytic shutdown” as a result of various imbalances of the hemostaseological homeostasis.^36,37  The elevated D-dimer levels in combination with the elevated fibrinogen concentrations and the results of the viscoelastic analysis may indicate a greater fibrinolysis resistance, an interpretation that is consistent with recently published research.^30,38  These alterations could contribute to the observed laboratory changes consistent with disseminated intravascular coagulation, but distinctively lead to thromboembolic events in these patients.

Further investigations and the determination of parameters representing the complex system of fibrinolysis (e.g., plasmin, plasminogen activator inhibitor, or thrombin-activated fibrinolysis inhibitor) might provide further insights into this topic. To gain definitive answers about the pathophysiology of coagulation in COVID-19, data from a larger patient population are warranted.

The major threat to internal validity results from a sampling bias of the study population: for the COVID-19 cohort, we included a substantial number of patients referred to our hospital by primary care providers. This may have resulted in an assessment of patients more severely affected by COVID-19 compared to the general COVID-19 population. This sampling bias is further substantiated by the fact that, in contrast to the overall COVID-19 population, all COVID-19 patients of this study were mechanically ventilated. The analysis of these highly selected, severely affected patients may result in an overestimation of the results observed within our study. Differential misclassification may result from the sedation/intubation of COVID-19 patients, leading to uncertainties regarding the patient history when compared to fully awake, healthy controls. We have demonstrated significant differences between both healthy controls and the COVID-19 cohort, which might serve as confounding variables and potentially interfere with the independent variable (diagnosis of COVID-19). However, little is known about the effects of this novel disease, and matching of a cohort by age, body mass index, and renal replacement therapy would both fail to reflect the complexity of this disease and leave other confounders unaddressed (such as diabetes, among others). Our stratified or bivariable analysis showed only minor differences or no influence of the confounders sex, age, and body mass index on the analyzed coagulation alterations from point-of-care diagnostics. At this point, we sought to provide a comparison to a healthy control group with all the limitations inferred by such a comparison, but we carefully avoid overclaiming our findings. However, we would like to stress that some studies on rotational thromboelastometry^39,40  have shown significant changes for a higher age toward hypercoagulability, and therefore, the age difference between the groups may have influenced our results. Moreover, an influence of obesity^41,42  and diabetes mellitus^43–45  has also been previously described and may have affected our results reported. Due to the novelty of the assay, no formal analysis of measures of reliability or validity has been published for the ClotPro yet. To avoid confusion regarding the maximum lysis results, which might appear contradictory to the lysis time results, it has to be considered that for technical reasons, the lowest amplitude of the maximum lysis is 1.5 mm during lysis, and therefore, the maximum lysis cannot reach 100%.

While we consider the internal validity of this study to be high regarding the values obtained by both laboratory testing and thromboelastometry, limitations arise from the nonlongitudinal nature of the measurements obtained for this study. Moreover, we assume that a sequential analysis of platelet function may better characterize alterations than a single analysis as performed in our study, since platelet function is a very dynamic process and blood sampling for impedance aggregometry was done in (mean) 7 ± 3.5 days after ICU admission. Therefore, increased platelet aggregation at an early phase of COVID-19 with subsequent exhaustion of platelets cannot be excluded by this study—particularly since the AUC of the ADP test was significantly decreased compared to the control group. While we provide a detailed analysis of the fibrinolysis resistance at one point of the disease, future studies should provide longitudinal information on the time point associated with the named pathology. We hypothesize that the findings of our study might be generalizable to the general COVID-19 population with the limitation that our findings were obtained in a group of severely affected, ventilated patients. Other manuscripts have been published on point-of-care diagnostics in COVID-19 demonstrating results similar to the findings of our study. However, future studies repeating measurements on thromboelastometry in patients with COVID-19 are warranted to replicate our findings and to further improve this study’s external validity.

Although critically ill patients with COVID-19 have a hypercoagulable state, we did not find greater platelet aggregability based on impedance aggregometric testing. Moreover, thromboelastometry in our patients revealed greater fibrinolysis resistance. These findings may contribute to the understanding of the hypercoagulable state in critically ill patients with COVID-19 and be used to further develop appropriate anticoagulation regimens for the prevention and treatment of thromboembolic events.

Support was provided solely from institutional and/or departmental sources.

Dr. Zacharowski’s department has received unrestricted educational grants from B. Braun Melsungen AG (Melsungen, Germany), Fresenius Kabi GmbH (Bad Homburg v.d.H., Germany), CSL Behring GmbH (King of Prussia, Pennsylvania), and Vifor Pharma GmbH (Glattbrugg, Switzerland). In the past 3 yr, Dr. Zacharowski has received honoraria or travel support for consulting or lecturing from the following companies: Abbott GmbH & Co KG (Wiesbaden, Germany), Aesculap Akademie GmbH (Melsungen, Germany), AQAI GmbH (Mainz, Germany), Astellas Pharma GmbH (Tokyo, Japan), Astra Zeneca GmbH (Wedel, Germany), Aventis Pharma GmbH (Frankfurt, Germany), B. Braun Melsungen AG, Baxter Deutschland GmbH (Unterschleißheim, Germany), Biosyn GmbH (Fellbach, Germany), Biotest AG (Dreieich, Frankfurt), Bristol-Myers Squibb GmbH (Munic, Germany), CSL Behring GmbH, Dr. F. Köhler Chemie GmbH (Bensheim, Germany), Draëger Medical GmbH (Luebeck, German), Essex Pharma GmbH (Haar, Germany), Fresenius Kabi GmbH, Fresenius Medical Care (Bad Homburg v.d.H., Germany), Gambro Hospal GmbH (Unterschleißheim, Germany), Gilead (Foster City, California), GlaxoSmithKline GmbH (Brentford, United Kingdom), Gruenenthal GmbH (Aachen, Germany), Hamilton Medical AG (Bonaduz, Switzerland), HCCM Consulting GmbH (Bremen, Germany), Loëwenstein Medical GmbH (Bad Ems, Germany), Janssen-Cilag GmbH (Neuss, Germany), med update GmbH (Wiesbaden, Germany), Medivance EU B.V. (Den Haag, The Netherlands), MSD Sharp & Dohme GmbH (Haar, Germany), Novartis Pharma GmbH (Basel, Switzerland), Novo Nordisk Pharma GmbH (Mainz, Germany), P.J. Dahlhausen & Co. GmbH (Cologne, Germany), Pfizer Pharma GmbH (Karlsruhe, Germany), Pulsion Medical Systems S.E. (Munic, Germany), Siemens Healthcare (Erlangen, Germany), Teflex Medical GmbH (Wayne, Pennsylvania), Teva GmbH (Mörfelden-Walldorf, Germany), TopMedMedizintechnik GmbH (Pressbaum, Austria), Verathon Medical (Bothel, Washington), and Vifor Pharma GmbH (Glattbrugg, Switzerland). Dr. Weber received honoraria for scientific lectures from CSL Behring GmbH, Haemonetics (Boston, Massachusetts), Biotest AG, Verum Diagnostica GmbH (München, Germany), enicor GmbH (München, Germany), IL Werfen (Barcelona, Spain), and Roche Diagnostics (Basel, Switzerland). The remaining authors declare no competing interests.