Critically ill patients represent a population at high risk of venous thromboembolism (VTE), which includes deep vein thrombosis (DVT) and pulmonary embolism (PE). These events are associated with high morbidity and mortality, with PE mortality reaching 65% in case of cardiac arrest [1]. Thromboprophylaxis is therefore recommended in critically ill patients [2, 3], with low-molecular-weight heparins being the most commonly used anticoagulants and probably associated with the best benefit–risk ratio [4]. The role of direct oral anticoagulants remains debated [5].

There are several particularities in critically ill patients. First, the usual warning signs of DVT and PE such as pain or dyspnea are most often missed in a critical care context in which patients are usually mechanically ventilated and sedated. In addition, in case of overt hemodynamic/respiratory instability, the patients might not be easily transportable, which may decrease the possibility of using certain diagnostic devices. Then, critically ill patients have a high risk of bleeding complications due to organ failure and multiple invasive procedures. Simultaneously, they also have an increased probability of developing venous thrombosis in unusual locations due to their pro-inflammatory status and the use of intravascular devices such as central venous or dialysis catheters, exposing these patients to the risk of catheter-related venous thrombosis.

Catheter-related venous thrombosis poses several issues. The first is that diagnosis can sometimes be difficult, especially in cases of progressive forms or in case of subclavian location of the catheter where the reliability of duplex of ultrasound for catheter-related thrombosis might lead to 21% of false-negative diagnoses by ultrasound [6]. The second is that the clinical relevance of these thromboses can vary from asymptomatic non-threatening events (thrombosis around the catheter) to potentially deadly PEs or superior vena cava syndrome. Finally, the epidemiology of catheter-related thrombosis is relatively well known in patients with cancer, but data remain scarce in critically ill patients [7].

Therefore, Wu et al. [8] should be thanked for conducting and publishing a large multicenter prospective study systematically evaluating the risk of catheter-related thrombosis in critically ill patients. More than 1200 patients were included and underwent a systematic screening for catheter-related thrombosis using Doppler ultrasound performed by an investigator trained in the diagnosis of venous thrombosis. The main result is that catheter-related thrombosis is frequent, occurring in 16.9% (95% confidence interval 14.8–18.9%) of patients. Several elements should be noted. First, while the overall rate was high, only a third were extensive thromboses, and none was symptomatic, although patients with catheter-related thrombosis experienced prolonged intensive care unit (ICU) stay. Half of the events were detected within the first 4 days post-venous puncture and 12% occurred on the 1st day, while only 7% of the patients had already received prophylactic anticoagulation. Catheter removal was associated with thrombosis resolution in all the patients in whom catheter was removed because of thrombosis, representing 14% of the patients with catheter-related venous thrombosis. Finally, only the internal jugular vein site and age were associated with the risk of thrombosis in multivariate analysis.

The results of this important study show a high frequency, higher than expected, of catheter thrombosis in ICU patients, which raises the question of the clinical relevance of these events. The evidence that would support the weight of events diagnosed during systematic screening is drawn from prevention strategies in surgery, particularly in orthopedic surgery, where thromboembolic events are identified through systematic ultrasound screening, as it was also done in this study. Conversely, it is possible to think that the severity of a catheter thrombosis could be different from distal thrombosis in orthopedic surgery, where the risk of migration is known, but not the risk of proximal extension and embolic migration.

Although the occurrence of a catheter thrombosis does not seem to worsen the patient prognosis (ICU mortality was not statistically different), it can significantly and lastingly alter venous accessibility in critically ill patients, one-third of the thromboses being extensive in the present cohort. However, no conclusion can be drawn on the anticoagulation management to be carried out in case of catheter-related venous thrombosis. Unfortunately, the therapeutic consequences of catheter-related thrombosis were not followed in this study and the decision of therapeutic anticoagulation was left to the discretion of the physician. Knowing the associated risk of therapeutic anticoagulation in critically ill patients, large randomized controlled trials investigating the need for therapeutic anticoagulation in case of catheter-related thrombosis are needed to answer this important issue.

Based on these data, the development of molecules targeting the intrinsic pathway of coagulation might open the way to strategies for preventing catheter-associated venous thrombosis [9]. The contact activation (intrinsic) pathway indeed plays a key role in thrombus growth, as thrombin amplifies the activation of factor XI, which will in turn amplify the downstream common pathway leading to thrombin and fibrin formation (Fig. 1). Targeting factor XI might thus allow to suppress the pathological thrombosis pathway. As factor XI might only play a minor role in physiological hemostasis, this pathway would remain largely unaffected, explaining why the bleeding risk might be lower than with other anticoagulant molecules.

Fig. 1
figure 1

Toward new therapeutics in thromboprophylaxis in critically ill patients? Both clot formation after tissue injury (physiological hemostasis) and clot formation resulting in vessel occlusion (pathological hemostasis—thrombosis) are initiated by tissue factor (TF). However, in the first process, FXI plays a minor role to seal off the injured vessel, while it is required for thrombus growth. Anticoagulants that could be used in thromboprophylaxis of critically ill patients inhibit FXa and/or thrombin (DOACs, heparins) with subsequent antithrombotic effects, but also significant bleeding risk by altering physiological hemostasis. Targeting FXI, which only plays a minor role in physiological hemostasis, might allow to reduce bleeding risk by preserving hemostasis while exerting an antithrombotic effect. DOACs direct oral anticoagulants, NETs neutrophil extracellular traps, PolyP polyphosphate, TF tissue factor

Full size image