Serial fibrin monomer and D-dimer plasma levels measurements can capture thrombotic complications in critically ill COVID-19 patients: A prospective observational study

Venous thromboembolism (VTE) is a frequent and dreaded complication in the intensive care unit (ICU), albeit often unrecognized, since its diagnosis is challenging in this setting. The use of biomarkers could be valuable for the differential diagnosis of VTE, and more broadly of thrombotic events.

D-dimer plasma level measurement belongs to the exclusion approach for suspected VTE in the outpatient setting, but is poorly effective in the inpatient setting [1]. Moreover, in critically ill patients, D-dimer levels are often far above the threshold for exclusion of VTE (e.g., 500 ng/mL), with very variable levels depending on the individual and the clinical context; the determination of a diagnostic threshold is therefore complicated.

The measurement of fibrin monomers (FMs) in plasma has been suggested in this setting, but still lacks firm validation [2], [3]. FMs are produced as a result of the action of thrombin on fibrinogen. They are highly reactive molecules that assemble to form protofibrils but can also associate with two fibrinogen molecules, which limits their molecular mass so that they remain soluble, and thus accessible to measurement in plasma [2]; at the same time such assembly prevents permeation of the extravascular space [4]. Such a specific marker of intravascular thrombin generation could overcome the limitations of D-dimer.

We previously reported a pilot study about the potential usefulness of longitudinal FMs measurements for the identification of thrombosis in critically ill coronavirus disease 2019 (COVID-19) patients, who represent a convenience cohort at high thrombotic risk [5]. We found that, in contrast to D-dimer, FMs plasma levels were mostly normal, but a sharp increase of them, the detection of which requires close monitoring, could be observed in presence of substantial intravascular thrombin generation leading to clinically significant fibrin deposits: VTE, arterial thrombosis, disseminated intravascular coagulation (DIC), clotting in an extracorporeal circuit. Those preliminary results deserved to be verified more robustly.

Hence we conducted a prospective, observational, monocenter study after approval by the ethics committee (NUB: B0392020000031). We included patients admitted consecutively to the ICU for severe COVID-19 (requiring either high-flow oxygen administration via nasal cannula or tracheal intubation) between October 1, 2020, and May 31, 2021. Patients were excluded if the ICU admission was for any reasons other than severe COVID-19, if the duration of ICU stay was less than seven days, or when citrated plasma samples were not available at least every two days.

Leftover plasma from citrated blood samples of included patients drawn at the request of the clinicians was frozen at −80 °C after double centrifugation (1500 g, 15 min, 20 °C). D-dimer and FMs were measured using the STA-Liatest D-Di Plus (results expressed as fibrinogen equivalent units) and STA-Liatest FM reagents, respectively, and the automated STA-R-Max 2 system (Stago).

Patients were managed according to local standards of care. During the study period, all patients received thromboprophylaxis according to the GIHP proposals [6] and corticosteroids (dexamethasone 6 mg once daily for 10 days). Doppler ultrasound of the lower limbs and computed tomography pulmonary angiogram (CTPA) were performed within 72 h of ICU admission.

The primary outcome was defined as any documented incident arterial or venous thrombotic event (i.e., compression ultrasonography for DVT or superficial vein thrombosis (SVT), CTPA, or echocardiography for PE, cerebral CT or MRI for stroke), or clotting in an extracorporeal circuit; thereafter collectively referred to as thrombotic events (TEs). DVT and PE diagnosed during the systematic admission workup were not included in the analysis. DIC was diagnosed according to the ISTH criteria.

Logistic mixed models were used to assess the association between the occurrence of a TE and D-dimer and FMs levels. Patients were censored after the TE or after 30 days of ICU stay. Models were fit by maximum likelihood and inference was performed by likelihood ratio tests. Alpha was set at 0.05, and all tests were two-sided.

We included 69 patients in the study (Fig. 1 ). During 1552 patients/days of follow-up a total of 1101 plasma samples were assayed. Median age was 63 years (IQR, 58–69) and median ICU stay was 21 days (IQR, 14–31) (Table S1). Twelve patients were on ECMO at some time-points of their ICU stay and 19 eventually died. Seventeen patients presented an asymptomatic DVT at admission screening. Over the ICU stay, 19 patients developed a TE, including 16 VTE (three PE with DVT, 11 isolated DVT, two SVT), one DIC, one myocardial infarction and one ECMO oxygenator clotting (Table S2). No patients experienced a recurrence of a TE.

Fig. 1.

Patients flow diagram.

ICU, intensive care unit; SARS-CoV-2, severe acute respiratory syndrome corona virus 2; RT-PCR, reverse transcriptase polymerase chain reaction; COVID-19, coronavirus disease 2019.

Median plasma D-dimer and FMs levels outside of TEs were 1900 ng/mL (IQR, 1225–2750) and 3.6 μg/mL (IQR, 3.0–4.6), respectively. The time-courses of D-dimer and FMs levels from ten days before the TE to ten days after are shown in Fig. 2 . Both D-dimer and FMs levels were associated with TEs (p < 0.001 for both). For example, an increase of 5000 ng/mL in D-dimer levels or an increase of 50 μg/mL in FMs levels were associated with an increased odd of TE of 3.9 (95 % CI, 1.8–8.4) or 5.6 (95 % CI, 2.7–12.0), respectively.

Fig. 2.

Time-related changes in D-dimer and fibrin monomers levels around the day of the incident thrombotic event (19 patients).

Day zero is defined as the day of the thrombotic event (either the beginning of clinical suspicion if the event was symptomatic, or the day of diagnosis if not). The median (IQR) D-dimer or fibrin monomer levels of the 19 patients who experienced an incident thrombotic event during their ICU stay are shown in the figure. The dashed lines correspond to the 500 ng/mL threshold for D-dimer and to the upper limit of the normal range (6 μg/mL) for fibrin monomers.

Those results evidenced that serial, closely monitored D-dimer or FMs plasma levels in critically ill COVID-19 patients can point to the occurrence of thrombotic events, and is in line with the results of our pilot study [5]. The advantage of FMs is that levels outside of TEs were close to the reference values in the majority of patients, making the definition of an alert threshold simpler than for D-dimer, the levels of which were much more elevated and variable. Indeed, plasma D-dimer can originate from both intravascular and extravascular fibrin deposition and breakdown, since their molecular mass can be small enough to permit transits between both compartments. In contrast, the source FMs, which have a higher molecular mass, is probably restricted to the intravascular compartment, and FMs are thus more specific to (intravascular) thrombotic disorders [7].

On the other hand, FM levels return to normal more rapidly (average half-lives of 2–6 h vs. >8 h for D-dimer) [4], especially if the patient is on anticoagulant therapy [8], [9]. This can lead to missing the peak when blood collection is not concomitant with or close to the acute phase of thrombosis and intravascular thrombin generation, i.e. during a potentially short window. Indeed, FMs levels were increased only for 24–48 h in some patients we studied; the implementation of enhanced-dose thromboprophylaxis and the rapid initiation of therapeutic anticoagulation after identification of a TE might have played a role in such transience. Regarding elevated D-dimer or FMs levels without identified TEs, potential causes could include unidentified asymptomatic TEs (clinical relevance of which is then questionable), inadequately anticoagulated patients, or preanalytical artifacts.

Regarding FMs, most studies agree on their ability to identify incident TEs in both the out and inpatient settings [8], [9], [10], [11], [12], [13], [14], [15], [16], [17], [18], [19], [20], [21], [22], [23], [24], [25]. Interestingly, the studies suggesting better accuracy of FMs over D-dimer for thrombosis diagnosis were those that iteratively measured both biomarkers around the TE [8], [9], [12], [22], [24]. In the COVID-19 setting, two studies failed to identify an association between FMs levels at admission and the occurrence of VTE [26], [27]; this is however not surprising as the biomarker would only increase when the thrombotic process begins. Two other studies, which serially measured both FRMs along the ICU stay of COVID-19 patients, concluded that both biomarkers could be suitable to identify TEs [5], [28].

The strength of the study resides in the very regular and close monitoring of the two biomarkers (at least one measurement every 48 h) and the total number of samples analyzed (>1000), enabling a close evaluation of their time-courses and an analysis of the data with a robust longitudinal statistical method. However, the number of patients remains low. Precise dating of TEs was not always possible; in fact, for three patients, D-dimer and FMs peaks preceded the thrombosis diagnosis and both biomarkers levels had already returned to baseline at diagnosis. Third, we were unable to implement routine and systematic screening for thrombosis during ICU stay. Asymptomatic thrombosis could have been missed that would explain some elevations in D-dimer or FMs. Finally, the study included only COVID-19 patients due to convenience (high incidence of TEs) and results deserve to be extended in a more general population of patients at high thrombotic risk.

To conclude, both D-dimer and FMs are capable of capturing incident TEs. D-dimer measurement has the drawback of being frequently elevated in inpatients, the determination of a diagnostic threshold being complicated in this setting. Consideration of their time-related changes (i.e., important increase within a few days) is therefore more appropriate. FMs have the advantage of being specific to thrombin generation within the intravascular compartment, resulting in levels close to normal values (i.e., 6 μg/mL) in patients who do not develop TEs. However, FMs have a shorter half-life than D-dimer, sometimes being increased for only 24 h, with an ensuing risk of missing their peak in case of measurement outside the very acute phase of intravascular thrombin generation. Notwithstanding, their usefulness in the work-up of a suspected TE in high-risk critically ill patients should be further studied.

Declaration of competing interest

The authors declare the following financial interests/personal relationships which may be considered as potential competing interests:

MH, IM, PB, BS and BN declare no conflict of interest.

EDM and TL report institutional fees from Stago.

AG reports personal fees from BMS-Pfizer and Sanofi for conference attendance, outside the submitted work.

FM reports institutional fees from Stago, Werfen, Nodia, Roche Sysmex and Bayer. He also reports speaker fees from Boehringer-Ingelheim, Bayer Healthcare, Bristol-Myers Squibb-Pfizer, Stago, Sysmex and Aspen all outside the submitted work.

Appendix A. Supplementary data

Supplementary tables