Applications of rotational thromboelastometry in heparin monitoring in critical COVID-19 disease: Observations in the Maastricht Intensive Care COVID cohort

Abstract

Background

Critically ill COVID-19 patients are at risk for venous thromboembolism (VTE). Therefore, they receive thromboprophylaxis and, when appropriate, therapeutic unfractionated heparin (UFH) or low-molecular-weight heparin (LMWH). To monitor heparins in COVID-19 disease, whole-blood rotational thromboelastometry (ROTEM) may be a promising alternative to the aPTT and anti-Xa assays.

Objective

To evaluate the ROTEM INTEM/HEPTEM ratios in mechanically ventilated COVID-19 patients treated with UFH and therapeutic LMWH.

Material and methods

A subcohort of mechanically ventilated COVID-19 patients of the prospective Maastricht Intensive Care Covid (MaastrICCht) cohort was studied. Anti-Xa, aPTT, and ROTEM measurements following treatment with UFH or therapeutic dose of LMWH (nadroparin) were evaluated using uni- and multivariable linear regression analysis and receiver operating characteristics.

Results

A total of 98 patients were included, of which 82 were treated with UFH and 16 with therapeutic LMWH. ROTEM-measured INTEM/HEPTEM CT ratio was higher in patients using UFH (1.4 [1.3–1.4]) compared to patients treated with LMWH (1.0 [1.0–1.1], p < 0.001). Both the aPTT and anti-Xa were associated with the CT ratio. However, the β-regression coefficient (95%CI) was significantly higher in patients on UFH (0.31 (0.001–0.62)) compared to therapeutic LMWH (0.09 (0.05–0.13)) for comparison with the anti-Xa assay. Furthermore, ROC analysis demonstrated an area under the curve for detecting UFH of 0.936(0.849–1.00), 0.851(0.702–1.000), and 0.645(0.465–0.826) for the CT ratio, aPTT, and anti-Xa, respectively.

Conclusion

The ROTEM INTEM/HEPTEM CT ratio appears a promising tool to guide anticoagulant therapy in ICU patients with COVID-19 disease, but associations with clinical endpoints are currently lacking.

1. Introduction

Critically ill patients with COVID-19 are known to be at high risk for venous thromboembolism (VTE) due to a hypercoagulable and inflammatory state [[1], [2], [3]]. Higher low molecular weight heparin (LMWH) dosages (so-called intermediate dosage) were initiated as thromboprophylaxis due to the frequent occurrence of VTE in critically ill COVID-19 patients [4]. Therapeutic dosages of LMWH are required in the Intensive Care Unit (ICU) for several indications, including VTE. Alternatively, indications for unfractionated heparin (UFH) include extracorporeal membrane oxygenation (ECMO) and continuous renal replacement therapy (CRRT). Most critically-ill COVID-19 patients treated with UFH present with heparin resistance [5], defined as requiring >35,000 IU per day to reach target-activated partial thromboplastin time (aPTT) or anti-Xa levels [6].

Patients treated with UFH require monitoring using the aPTT or anti-Xa laboratory assay. In most laboratories, the aPTT is the conventional assay used for UFH monitoring. Although aPTT is a simple, relatively cheap, and widely available test, it is known to be poorly standardized and affected by both preanalytical and analytical parameters unrelated to heparin anticoagulant activity [7]. E.g., UFH-mediated aPTT prolongation is highly dependent on the reagent and analyzer used [8]. LMWH does not generally require monitoring, but in vitro levels may be desired upon clinical indication, such as developing VTE under prophylaxis. LMWH is monitored using the anti-Xa assay, although this class of heparins can prolong aPTT depending on their molecular size distribution, anti-IIa activity, and timing of blood withdrawal (peak or trough) [9]. However, both assays have limited applicability to evaluate the global hemostatic status of a patient. In vitro monitoring of UFH in patients with severe COVID-19 is challenging, as therapeutic ranges for aPTT and anti-Xa are discordant, and no gold standard is available [5].

To better evaluate the in vitro heparin effect, it was proposed to monitor both anti-Xa and aPTT in patients with heparin resistance [10], which was common in ICU Covid-19 patients [11], increasing workload and complicating interpretation when discordance occurs. Therefore, it was hypothesized that a global, whole-blood assay, such as rotational thromboelastometry (ROTEM), may better illustrate the in vivo effect of heparins in COVID-19 patients admitted to the ICU [12].According to the manufacturer, ROTEM can be applied in cardiac surgery to detect the presence of heparin by comparing the clotting time (CT) of the INTEM (intrinsic pathway) and HEPTEM (INTEM with heparinase to neutralize the heparin effect) assays [13]. In the absence of UFH, it is assumed that the INTEM and HEPTEM tests produce a similar CT. It was previously demonstrated that the ROTEM INTEM/HEPTEM CT ratio could detect patients at risk for bleeding under heparin [13]. ROTEM is a whole blood test and may therefore better reflect the in vivo heparin effect in patients with severe COVID-19, characterized by elevated inflammation parameters and hypercoagulability [2]. The objective of this observational study is to compare ROTEM INTEM/HEPTEM ratios, aPTT, and anti-Xa and to evaluate ROTEM's applicability to detect the presence of heparins in mechanically ventilated COVID-19 patients treated with UFH and therapeutic LMWH.

2. Material and methods

The manuscript was written following the STrengthening the Reporting of OBservational studies in Epidemiology (STROBE) guidelines [14].

2.1. Patients and study design

The Maastricht Intensive Care COVID (MaastrICCht) cohort is a prospective study conducted on patients admitted to the ICU of the Maastricht University Medical Centre (Maastricht UMC+) [15]. All included patients were mechanically ventilated and had signs and symptoms of a viral infection, including a PCR positive for SARS-CoV-2 and/or a chest CT scan, scored positive based on a CORADS-score of 4–5 [16]. Patients were admitted from the MUMC + emergency department, non-ICU wards, or transferred from other ICUs either for tertiary care referral, such as ECMO, or due to lack of bed availability in regional hospitals.

Every day, a comprehensive and uniform set of clinical, physiological, and laboratory variables was collected.

ECMO consisted of venovenous extracorporeal membrane oxygenation (VV-ECMO) with the permanent life support (PLS) and Cardiohelp heart-lung support (HLS) system (Maquet Cardiopulmonary, Rastatt, Germany). Indications for VV-ECMO support were a P/F-ratio <80 mmHg despite prone positioning and optimal PEEP, ventilator support <6 days, age <70 years, mono-organ failure, and no severe comorbidities. CRRT was initiated in case of acute kidney insufficiency (AKI) KDIGO stage 3 and consisted of continuous venovenous hemodiafiltration (CVVHD) using the multiFiltrate system (Fresenius Medical Care, Bad Homburg, Germany). Initially during the pandemic, due to the citrate shortage, UFH was used for anticoagulation according to local procedures.

The local institutional review board (Medisch Ethische Toetsingscommissie (METC) 2020–1565/300523) of the Maastricht UMC + waived consent and approved the MaastrICCht cohort study in accordance with the Declaration of Helsinki. Patients or their families consented to participate.

2.2. Blood collection, preparation, and storage

Daily blood samples from all patients were collected from an arterial line using 7.2 mg K2 EDTA (4.0 mL), serum, and 3.2% (w/v) citrate Vacutainer vacuum blood tubes (Becton Dickinson, Plymouth, UK). Additional 3.2% (w/v) citrate blood tubes were collected twice weekly for global assays of hemostasis, including ROTEM. Patient inclusion in the MaastrICCht cohort started from March 25^th 2020 and enrolment for the hemostasis subcohort started later, on April 23^rd 2020 [15]. From January 19th, 2021, ROTEM measurements were reduced to once weekly. Whole blood tests were performed within 4 h after blood withdrawal. Platelet-free plasma (PFP) was obtained using two subsequent centrifugation steps: initial centrifugation of 2490 g for 5 min, followed by 10,000 g for 10 min.

2.3. Laboratory measurements

The aPTT (Dade Actin FSL; Siemens, Marburg, Germany) and anti-Xa (Biophen Heparin LRT; Hyphen Biomed, Neuville-Sur-Oise, France) were performed on a Sysmex CS2100i (Sysmex Corporation, Kobe, Hyogo, Japan) hemostasis analyzer in 3.2% citrated blood. Local aPTT reference values were 23–32s. For the anti-Xa measurement, samples of COVID-19 patients (18 μl) were diluted three times with reference pooled plasma (36 μl). Anti-Xa activity was determined using an LMWH calibration line (anti-Xa-LMWH; Hyphen Biomed).

2.4. Rotational thromboelastometry (ROTEM)

ROTEM (Werfen; Barcelona, Spain) is a point-of-care viscoelastic assay performed in whole blood. The ROTEM sigma cartridge-based system was performed to evaluate the EXTEM, FIBTEM, INTEM, and HEPTEM assays. The EXTEM and FIBTEM results are discussed elsewhere [3]. The INTEM assay is activated using ellagic acid and evaluates the intrinsic coagulation pathway (resembling aPTT). The HEPTEM assay is similar to the INTEM assay but includes heparinase, which neutralizes the anticoagulant effect of heparins if present. The clotting time (CT) is the period from the start of an assay until a clot amplitude of 2 mm is achieved, which illustrates the time until the initiation of clot formation. Therefore, INTEM, but not HEPTEM CT, becomes prolonged in the presence of heparins. The INTEM/HEPTEM CT ratio is calculated by dividing the INTEM CT by the HEPTEM CT and a ratio >1 indicates the presence of heparin anticoagulant activity. Similar calculations can be performed for the clot firmness time (CFT) and maximum clot firmness (MCF).

2.5. Anticoagulation

Indications for UFH were CRRT or ECMO. UFH was monitored using aPTT, with a heparin therapeutic range (HTR) of 50–80 s, based on an aPTT ratio of 1.5–2.5 fold baseline, and adjustments of heparin dosage at the bedside were made following aPTT according to local guidelines. In addition, anti-Xa was measured, but results were not used for changes in UFH dosage. The LMWH of choice was nadroparin. Patients received therapeutic LMWH when diagnosed with VTE (both deep venous thrombosis and pulmonary embolism) or when at risk for cardiac embolic complications following national guidelines [17]. Computed tomography pulmonary angiography (CTPA) was used as standard for detection of clinically suspected pulmonary embolism. Therapeutic nadroparin dosages were defined as twice daily 3,800, 5,700, and 7600 IU for, respectively, a weight of <50, 50–70 kg, and 70–110 kg. Patients received nadroparin between 8:00–10:00 a.m. and 8:00–10:00 PM.

In the current analysis, we selected and studied all patients with a ROTEM measurement while being treated with UFH and therapeutic nadroparin from the start of the MaastrICCht until March 23^th 2021. Anticoagulant use on the day prior to blood withdrawal was used to classify patients in the UFH or therapeutic LMWH group. As blood withdrawal took place at 5:00 a.m., the anticoagulant strategy generally remained unchanged over night.

2.6. Statistical analysis

The first available ROTEM measurement after initiation of therapeutic LMWH or UFH was included in the analysis. Population characteristics were described using mean (standard deviation; SD) or median [Interquartile range; IQR]. Population characteristics and hemostatic assays were compared using independent Student's t-test or Mann-Whitney U test as appropriate.

Correlations between ROTEM measurements and platelet count were assessed with Spearman correlation coefficients (ρ). The cross-sectional associations between aPTT and anti-Xa with ROTEM INTEM/HEPTEM CT, CFT and MCF ratios were investigated using linear regression analyses. We investigated the association between the coagulation assays (aPTT and anti-Xa) and the INTEM/HEPTEM ratio (model 1). Furthermore, the heparin class (UFH use yes/no) itself and the interaction between the heparin class and coagulation assay (either aPTT or anti-Xa) were added to model 1 (model 2). When the coagulation assay*UFH interaction term was statistically significant, the beta-coefficient was calculated separately (i.e., stratified) for the UFH and LMWH patient groups. The discriminating ability of ROTEM, aPTT, and anti-Xa to identify patients on UFH was assessed using receiver operating characteristics (ROC) analyses with the calculation of the area under the curve (AUC). Furthermore, sensitivity and specificity for the detection of UFH were evaluated for the ROTEM INTEM/HEPTEM CT ratio using the manufacturer's cut-off of 1.1.

A p-value of <0.05 and a p-value for interaction of <0.1 were considered statistically significant. Statistical analyses were performed in IBM SPSS Statistics for Windows, version 28.0 (IBM Corp., Armonk, N·Y., USA).

3. Results

Of the 232 patients included in the MaastrICCht cohort between March 25^th 2020 and March 23^th 2021, 160 were included in the hemostasis subcohort, of whom ROTEM measurements were available for 82 patients on therapeutic dose LMWH (nadroparin) and 16 patients on therapeutic dose UFH (Fig. 1 ). Two patients received both therapeutic LMWH and UFH throughout ICU admission and were classified within the UFH group. Patients with a ROTEM on prophylactic or intermediate dosage LMWH only were excluded (n = 60). We thus report on 98 patients. Median [IQR] age was lower in the UFH group (60 [51–67] years) compared to patients on therapeutic dose LMWH (66 [60–73] years), p-value = 0.01 (Table 1 ).

Fig. 1.

Flowchart

UFH = unfractionated heparin, LMWH = low molecular weight heparin, ROTEM = rotational thromboelastometry.

Table 1.

Population characteristics.

Variables	UFH	Therapeutic LMWH	p-value
Number of patients	16	82
Age (years)	60 [51–67]	66 [60–73]	0.01
Gender (male)	14 (87.5%)	69 (84.1%)
BMI (kg/m2)♢	26.7 [25.3–29.0]	27.8 [25.8–31.1]	0.19
Medical history
Chronic kidney disease, n (%)	2 (12.5%)	1 (1.2%)
Diabetes mellitus	2 (12.5%)	17 (20.7%)
Malignancy	1 (6.3%)	4 (4.9%)
Myocardial infarction	0	14 (17.1%)
Peripheral vascular disease	1 (6.3%)	4 (4.9%)
Medication prior to admission
Antiplatelet agent	2 (12.5%)	15 (18.4%)
DOAC total	2 (12.5%)	11 (13.4%)
Admission origin
•Emergency department	4 (25.0%)	13 (15.9%)
•Regular ward	3 (18.8%)	24 (29.3%)
•Transfer from other ICU	9 (56.3%)	45 (54.9%)
APACHE II at admission♢	14 [12–20]	13 [9–17]	0.31
SAPS II at admission♢	36 [30–42]	37 [29–42]	0.86
ICU stay
Length of stay ICU (days)	25 [20–42]	18 [11–33]	0.03
Time on ventilator (days)♢	27 [19–40]	17 [9–40]	0.06
CRRT during ICU stay (yes)	9 (56.3%)	3 (3.7%)
ECMO during ICU stay (yes)	9 (56.3%)	0
CT confirmed pulmonary embolism (yes)	7 (43.8%)	42 (51.2%)
Compression ultrasonography confirmed deep venous thrombosis (yes)	1 (6.3%)	3 (3.7%)
ICU mortality (death)	9 (56.3%)	35 (42.7%)

Data are means ± standard deviation (SD), median [interquartile range (IQR)], and n (%) with p-values tested by Mann-Whitney U test and student t-test as appropriate. ♢ Data are missing for BMI (therapeutic LMWH, n = 1), APACHE II, SAPS II (UFH group, n = 3, therapeutic LMWH group, n = 16), and ventilator days (therapeutic LMWH group, n = 28) due to transfers from other ICU's.

The INTEM/HEPTEM CT and CFT ratios were higher in patients using UFH compared to therapeutic dose LMWH (Table 2 , p-value <0.001). Similarly, the aPTT was prolonged in patients receiving UFH compared to the therapeutic dose LMWH. Spearman ρ demonstrated statistically significant correlations between INTEM parameters and platelet count, but not for the INTEM/HEPTEM ratios (Table 3 ). Linear regression analyses (Table 4 ) illustrated an association between anti-Xa and INTEM/HEPTEM CT ratio (β (95%CI): 0.15 (0.07–0.23)) and between aPTT and INTEM/HEPTEM CT ratio (0.008 (0.007–0.009)). We observed statistically significant interaction by heparin class in the association between anti-Xa and INTEM/HEPTEM CT ratio. Stratification by heparin class showed that the association between anti-Xa and INTEM/HEPTEM CT ratio in patients on UFH (β (95%CI) 0.31 (0.001–0.62)) was stronger compared to patients on LMWH (0.09 (0.05–0.13)). No interaction by heparin class was observed in the association between the aPTT and INTEM/HEPTEM CT ratio. Linear regression analyses for the INTEM/HEPTEM CFT and MCF ratios are presented in Supplementary Tables 1 and 2

Table 2.

INTEM/HEPTEM CT ratio, aPTT, and anti-Xa values for the first available ROTEM measurement after the start of unfractionated heparin (UFH) or therapeutic low molecular weight heparin (LMWH).

	UFH	No	Therapeutic LMWH	No	p-value
Days after intubation	10 [7–19]	16	8 [5–17]	82
ROTEM
INTEM CT (in s)	256 [239–286]	16	188 [175–205]	82	<0.001
INTEM CFT (in s)	58 [39–66]	16	48 [39–58]	82	0.12
INTEM MCF (in mm)	74 [70–79]	16	73 [69–77]	82	0.38
HEPTEM CT (in s)	183 [176–204]	16	184 [171–202]	82	0.73
HEPTEM CFT (in s)	46 [39–56]	16	50 [40–61]	82	0.38
HEPTEM MCF (in mm)	74 [70–78]	16	71 [67–75]	82	0.07
INTEM/HEPTEM CT ratio	1.4 [1.3–1.4]	16	1.0 [1.0–1.1]	82	<0.001
>1.1	14 (87.5%)		11 (13.4%)
≤1.1	2 (12.5%)		71 (86.6%)
INTEM/HEPTEM CFT ratio	1.2 [1.1–1.3]	16	0.97 [0.90–1.0]	82	<0.001
INTEM/HEPTEM MCF ratio	1.00 [0.99–1.01]	16	1.02 [1.01–1.04]	82	<0.001
Routine laboratory parameters
aPTT (in s)	65 [46–74]	16	30 [28–34]	76	<0.001
>50a	12 (75%)		0 (0%)
<50	4 (25%)		76 (100%)
anti-Xa (in IU/mL)	0.71 [0.58–0.86]	10	0.55 [0.44–0.74]	53	0.15
Platelet count (in 109/L)	248 [173–385]	16	317 [222–382]	78	0.384

Data are presented as mean (SD), median [IQR,] or n(%). No denotes the number of measurements.

^aBased on an aPTT heparin therapeutic range of 50–80s as applied in our hospital.

Table 3.

Spearman ρ between ROTEM parameters and platelet count stratified by UFH and therapeutic LMWH.

ROTEM parameters	UFH			Therapeutic LMWH
	Spearman ρ	P-value	No	Spearman ρ	P-value	No
INTEM CT	−0.531	0.034	16	−0.264	0.019	78
INTEM CFT	−0.760	<0.001	16	−0.562	<0.001	78
INTEM MCF	0.796	<0.001	16	0.690	<0.001	78
INTEM/HEPTEM CT ratio	−0.247	0.36	16	−0.173	0.13	78
INTEM/HEPTEM CFT ratio	−0.247	0.36	16	−0.173	0.13	78
INTEM/HEPTEM MCF ratio	−0.156	0.56	16	−0.172	0.13	78

Table 4.

Associations between coagulation assays (aPTT and anti-Xa) and the INTEM/HEPTEM CT ratio and interaction with UFH/LMWH. When applicable, stratified analyses were performed on patients using UFH and LMWH separately.

Dependent variable: INTEM/HEPTEM CT ratio	B (95% CI)	p-value
Independent variable: anti-Xa
Model 1		<0.001
anti-Xa	0.15 (0.07; 0.23)	<0.001
Model 2		<0.001
anti-Xa	0.09 (0.04; 0.15)	<0.001
UFH use (yes/no)	0.05 (−0.06; 0.17)	0.376
Interaction anti-Xa *UFH use	0.22 (0.07; 0.37)	0.005
Independent variable: aPTT
Model 1		<0.001
aPTT	0.008 (0.007; 0.009)	<0.001
Model 2		<0.001
aPTT	0.005 (0.001; 0.008)	0.016
UFH use (yes/no)	0.198 (0.022; 0.374)	0.028
Interaction aPTT*UFH use	0.000 (−0.004; 0.004)	0.91
Stratified results: anti-Xa
UFH patients	0.31 (0.001–0.62)	0.050
LMWH patients	0.09 (0.05–0.13)	<0.001

Data are β regression coefficients with 95% confidence intervals and p-values. Model 1: crude. Model 2: model 1 with an interaction term.

ROC analysis illustrated an area under the curve (AUC) for the detection of UFH of 0.936, 0.896, 0.795, 0.851, and 0.645 for the INTEM/HEPTEM CT ratio, CFT ratio, MCF ratio, aPTT, and anti-Xa, respectively (Table 5 ). By applying the manufacturer's cut-off of 1.1 for the detection of UFH, the sensitivity and specificity of the INTEM/HEPTEM CT ratio were 87.5% and 86.6%, respectively.

Table 5.

ROC analysis for the detection of UFH.

Variable	Number of measurements	AUC (95%CI)	p-value
INTEM/HEPTEM ratio	98	0.936 (0.849–1.000)	<0.001
INTEM/HEPTEM CFT ratio	98	0.896 (0.792–1.000)	<0.001
INTEM/HEPTEM MCF ratio	98	0.795 (0.693–0.898)	<0.001
aPTT	92	0.851 (0.702–1.000)	<0.001
anti-Xa	63	0.645 (0.465–0.826)	0.148

AUC = area under the curve.

4. Discussion

The present study had four main findings. First, we found that the INTEM/HEPTEM CT and CFT ratios could detect the presence of UFH compared to therapeutic LMWH. Second, platelet count demonstrated significant correlations for INTEM parameters, but not for the INTEM/HEPTEM ratios. Third, aPTT and anti-Xa were associated with the INTEM/HEPTEM CT ratio. The β-regression coefficient between anti-Xa and INTEM/HEPTEM CT ratio was higher in patients on UFH compared to patients on LMWH. This suggests that for every increase in anti-Xa, the INTEM/HEPTEM value rises more in patients on UFH, illustrating ROTEM's sensitivity for heparin's anticoagulant effect on IIa in addition to Xa. Last, ROC analysis demonstrated good discriminative ability for the presence of UFH compared to therapeutic LMWH.

Though most laboratories apply the aPTT assay, both the aPTT and anti-Xa assay are available for monitoring UFH in clinical practice [17]. When the aPTT is considered inaccurate or misleading, such as in patients with factor-deficiency states, disseminated intravascular coagulation, or antiphospholipid antibodies, the anti-Xa assay may better reflect UFH anticoagulation [11]. Though both assays are deemed appropriate for UFH monitoring, major discordance was observed between the aPTT and anti-Xa in several populations [18,19]. However, no differences in bleeding and thrombotic complications have been observed in a trial comparing UFH titration based on aPTT versus anti-Xa [20]. Recently, the discordance between aPTT and anti-Xa was also observed in a severe COVID-19 population on CRRT and ECMO [5]. Here, it was proposed that more global assays of hemostasis, such as ROTEM, may better illustrate the complex intricacies, among which heparin resistance is at play in COVID-19 patients. Heparin resistance is often defined as requiring >35.000 IU UFH in 24 h to reach target aPTT or anti-Xa levels [11]. A heparin therapeutic range of 0.3–0.7 IU/mL is widely accepted for the anti-Xa assay [21,22]. The heparin therapeutic range for the aPTT assay can be determined using a ratio of 1.5–2.5 (patient sample divided by a reference pool) or a range corresponding with the therapeutic anti-Xa levels [[23], [24], [25]]. Multiple mechanisms may be responsible for heparin resistance, including antithrombin deficiency, thrombocytosis, and binding of the negatively charged UFH molecules to a large variety of proteins in the circulation [11]. In COVID-19, elevated levels of coagulation factors (including fibrinogen) were considered to limit UFH-mediated aPTT prolongation [26]. This may explain the discordance with the anti-Xa assay, which is insensitive to an increase in coagulation factors. However, no association between the aPTT, fibrinogen, and inflammatory markers was observed in critically ill COVID-19 patients treated with UFH [5]. Of note, our results suggest that the heparin-anticoagulant effect as detected in the INTEM/HEPTEM ratios is independent of platelet count, contrarily to detection using the INTEM assay.

For LMWH, the anti-Xa assay is the laboratory assay of choice [17]. However, aPTT values also show prolongation depending on the inhibitory properties of the specific LMWH [27,28]. Shorter LMWH fragments lack the ability to accelerate the inactivation of thrombin by antithrombin [29]. Consequently, larger LMWH molecules are more likely to prolong coagulation assays, such as the aPTT. Nadroparin has a relatively short chain-length (4500 Da), therefore, predominantly exercises its anticoagulant effect by factor Xa inhibition [29]. This is illustrated by the limited prolongation of aPTT in the therapeutic LMWH group compared to local reference values. Due to differences in pharmacodynamic properties, other LMWHs may demonstrate different results. Previous studies applying ROTEM established that in vitro addition or in vivo administration of LMWH (nadroparin and Enoxaparin, respectively) results in limited CT prolongation [12,30]. Here, similar insensitivity to therapeutic nadroparin was observed.

The INTEM/HEPTEM CT ratio shows the largest median difference between patients treated with UFH versus therapeutic LMWH and is available as a fast, bedside result, when doing ROTEM. In comparison with routine tests. Therefore, this parameter is considered the most suitable for (potential) implementation in a clinical setting. The aPTT and anti-Xa assay were associated with a higher ROTEM INTEM/HEPTEM CT ratio. Noticeably, the relation between aPTT and ROTEM INTEM/HEPTEM CT ratio was independent of the heparin class, whereas the association with the anti-Xa assay was different for UFH and therapeutic dose LMWH. This suggests that for every point increase in anti-Xa, the INTEM/HEPTEM CT value rises more in patients on UFH compared to LMWH. The different assay characteristics may explain this observation. Specifically, the anti-Xa assay measures factor Xa inhibition only (main target of LMWH), whereas the INTEM/HEPTEM CT ratio and aPTT evaluate the intrinsic coagulation pathway, including factor Xa and IIa (main targets of UFH). Taken together, the ROTEM INTEM/HEPTEM CT ratio, similar to the aPTT, may better represent the in vivo effect of heparin. Furthermore, many preanalytical factors affecting the traditional assays can be disregarded due to the whole blood characteristic, standardized reagents, and cartridge-based ROTEM sigma system. Furthermore, the ROTEM has a quick turnaround time since no centrifugation is required, and the clotting time is generally achieved within minutes. UFH levels up to 7 IU/mL can be neutralized in the HEPTEM assay. Taken together, the ROTEM device may be of interest in the monitoring of UFH. However, clinical endpoints should be evaluated before clinical implementation of the ROTEM INTEM/HEPTEM CT ratio can be considered for UFH monitoring in an ICU setting.

Our study has several limitations. First, the sample size in the UFH group was too small to evaluate clinical outcomes, such as bleeding, thrombosis, and mortality, and the results cannot be generalized beyond mechanically ventilated ICU patients with COVID-19. Second, national guidelines support the use of UFH in the case of ECMO or CRRT in COVID-19, resulting in a more critically ill UFH group compared to patients on therapeutic dose LMWH [31]. However, as both the INTEM and HEPTEM assays are equally sensitive to hemostatic disruptions in COVID-19, it is unlikely that higher disease severity by itself significantly increases the ratio in the absence of heparins. Due to the relatively high prevalence of CRRT and ECMO in our cohort, specificity is likely to be overestimated compared to the general ICU population. Third, the time from the start of therapeutic LMWH or UFH and the ROTEM measurement was variable by design. No peak heparin levels were evaluated as blood withdrawal took place >4 h after LMWH administration. Withdrawal of ROTEM, anti-Xa and conventional coagulations tests took place at the same time, and we compared these results with ROTEM. Therefore we believe our results would be similar, when we would have measured peak LMWH levels.

Fourth, our study is cross-sectional by design, and by this method, no inferences about causality can be made. Fifth, we did not perform a complete diagnostic prediction study [32], and the observed sensitivity and specificity may vary when disease prevalence differs [33]. Future larger studies should evaluate the effect of other relevant laboratory parameters (e.g. antithrombin levels) on the association between ROTEM parameters and aPTT/anti-Xa.

5. Conclusion

The presented observations suggest a more in vivo representation of the UFH-anticoagulant effect using the ROTEM INTEM/HEPTEM CT ratio compared to the anti-Xa assay. In this respect, the whole-blood ROTEM assay may even proof to be superior compared to the platelet poor plasma based aPTT in future research. It should be noted that therapeutic LMWH was not detected using the CT ratio, limiting ROTEM's applicability to patients treated with UFH only. Taken together, the INTEM/HEPTEM CT ratio appears promising to guide anticoagulant therapy in ICU patients on UFH. However, future studies evaluating clinical endpoints in COVID-19 and other critically ill patients are required prior to implementation in clinical practice.

Declaration of competing interest

The authors declare the following financial interests/personal relationships which may be considered as potential competing interests: Renke H. Olie has received research support and honoraria from Bayer, Pfizer/BMS, Leo Pharma, Portola, and Sanofi. Hugo ten Cate received funding for research from Bayer and Pfizer, is a stakeholder in Coagulation Profile, is a consultant for Alveron, and has served on advisory boards for Bayer, Pfizer, Daiichi, Leo, and Gilead. Yvonne Henkens has received ROTEM cartridges free of charge for previous research unrelated to the current manuscript. All others report no conflicts of interest.

Handling Editor: Dr P Emmanouil

Appendix A. Supplementary data

The following are the Supplementary data to this article.