The Correlation of Hemostatic Parameters with the Development of Early Sepsis-Associated Encephalopathy. A Retrospective Observational Study

Sepsis is a global health problem, estimated in a 2017 report to affect almost 49 million people and to cause 11 million deaths, representing almost a fifth of all global deaths worldwide [1]. According to Sepsis-3 definition, sepsis is caused by a dysregulated immune response to infection, and diagnosed clinically by the development of new organ dysfunction due to infection [2]. Neurological dysfunction is one of the most common complications seen both in early and late stages of sepsis, affecting from one third to up to 70% of septic patients [3,4,5]. The neurological dysfunction in sepsis is known as sepsis-associated encephalopathy (SAE), a diffuse cerebral impairment caused by the host response to infection in the absence of a direct central nervous system (CNS) infection or other causes for encephalopathy [6, 7]. In a recent study including more than 4000 septic patients from the Medical Information Mart for Intensive Care IV (MIMIC-IV) and the eICU databases in the United States, SAE defined as a Glasgow Coma Scale score less than 15 or abnormal neurological findings consistent with delirium occurred in up to 68% of patients [8].

The spectrum of clinical manifestations of SAE is large, ranging from mild neurological dysfunction manifested as physiological and behavioral modifications to the most severe forms of delirium and coma [7, 9,10,11]. Due to the variable clinical manifestations, the diagnosis of SAE can be challenging especially in the early stages of sepsis, and the incidence, prevalence and mortality associated with SAE are difficult to assess [3, 4].

The pathophysiology and the complex mechanisms involved in SAE are still not completely elucidated. There are several factors involved, such as macro-and microcirculatory disturbances, dysfunction of the blood–brain barrier, neuro-inflammatory processes, or dysregulation of neurotransmitters [4, 12,13,14,15]. Until now, there is not any biomarker validated for the prediction or confirmation of SAE [4, 16, 17]. It is extremely important not to delay the diagnosis of neurological dysfunction, especially in patients with mild symptoms such as somnolence or behavioral changes, which sometimes represent the first clinical signs of organ dysfunction in sepsis. The identification of an organ dysfunction such as SAE leads to an earlier diagnosis of sepsis, allowing the detection of patients with highest morbidity and mortality and a quicker escalation of therapy [4, 18, 19].

It is known that the majority of patients with sepsis present different degrees of coagulation activation, ranging from mild coagulation disturbances to overt disseminated intravascular coagulation (DIC), associated with microcirculatory dysfunction, micro-thrombosis and organ dysfunctions [20, 21]. Cerebral hypoxic/ischemic lesions are commonly found in postmortem studies in patients with sepsis [12, 13, 22, 23]. These can be related to decreased cerebral flow and autoregulation failure, but also to hypercoagulability, microcirculatory dysfunction and endothelial damage [7, 13]. These coagulation changes leading to intravascular thrombosis and sometimes to simultaneous bleeding events are known and commonly seen in severe patients in later stages of sepsis, but the hemostatic changes associated with early development of organ dysfunction in sepsis are much less studied. Therefore, the aim of this study was to assess the coagulation profile in patients with early SAE and to compare the hemostatic parameters between septic patients with and without SAE in the first 24 hours from sepsis diagnosis.

This retrospective observational study included 280 patients with sepsis (52.1% men), with a mean (±SD) age of 69.64 (±15.27) years. 184 patients (65.7%) presented sepsis-associated encephalopathy (SAE) at study inclusion.

The parameters from the CBC, the age and gender of patients were compared between patients with and without SAE for the whole study group (n=280). The results are presented in Table 1. Following univariate analyses, the variables with unadjusted p values less than 0.1 were used for multivariate logistic regression analysis. Among the variables tested, age, leucocyte and platelet count were selected to be included in the multivariable logistic regression model (Table 2). After adjusting for confounders such as age and leucocyte counts, only the platelet count was useful to predict early SAE.

The hemostatic parameters from SCTs, plasmatic levels of coagulation factors and inhibitors, and ROTEM data were available in 118 patients. For both groups of patients (with and without SAE), the median values of prothrombin time were slightly increased compared to upper normal limit of the laboratory, while the mean plasma levels of both pro and anticoagulant factors were lower compared to the reference ranges provided by the laboratory (Table 3).

The data were compared between patients with and without early SAE (Table 3). The results showed a similar coagulation initiation in both groups of patients without any significant differences in PT values, plasmatic levels of pro- and anticoagulant factors, and clotting times.

The viscoelastic parameters reflecting clotting initiation were not different between the patients with and without SAE (Figure 1). However, the patients with early SAE presented a faster and more active coagulation propagation, with higher velocities of clot formation and shorter clot formation time on ROTEM (Figure 2, Table 3). The clot firmness, elasticity, and platelet component were also higher in patients with SAE compared to patients without SAE (Figure 3, Table 3). Clot lysis was not different between the groups. Although the plasmatic fibrinogen levels were not different between the groups, ROTEM analysis revealed a higher clot strength in FIBTEM in patients with than in patients without early SAE.

Fig. 1.

Comparison of coagulation initiation between groups.

Fig. 2.

Comparison of coagulation propagation between groups.

Fig. 3.

Comparison of clot strength between groups.

To compare the significance of platelet contribution or fibrinogen contribution to the overall increased clot strength that correlates with early SAE, plasmatic fibrinogen and the platelet component were selected to be included in a multivariable logistic regression model. The results show that platelet component remained a good predictor of SAE after adjusting for fibrinogen levels (Table 4).

The main result of this study is that the early development of SAE is correlated with a higher platelet count, with faster coagulation propagation phase, and increased clot strength. Our results suggest that patients with early SAE present with more important platelet contribution to clot formation and strength compared to patients without SAE.

Thrombocytopenia and plasmatic coagulation disturbances are frequent findings in sepsis, ranging from mild abnormalities to overt DIC in severe cases [21]. The severity of hemostatic alterations and of thrombocytopenia is usually correlated with the severity of the disease and the outcome of septic patients. However, this applies to later stages of sepsis, as in the early phases of sepsis the platelet count and coagulation parameters are usually maintained or even in the hypercoagulable range. In severe infections and sepsis, there is a bidirectional interaction between inflammation and coagulation, leading to an activated coagulation system together with dysfunctional anticoagulant and fibrinolytic pathways [31]. This initial coagulation activation is often not detected in clinical practice, but in later phases will lead to the consumption of coagulation factors and platelets, endothelial dysfunction and microvascular thrombosis which contributes to organ dysfunctions in severe cases [31, 32].

Rotational thromboelastometry (ROTEM) is a whole-blood point-of-care test useful for global assessment of hemostasis. It reflects the contribution of both plasmatic and cell-based coagulation during each phase of the coagulation process: initiation, propagation, clot formation and strength, and clot lysis[25]. Previous studies using ROTEM revealed a hypercoagulable profile in patients with early stages of sepsis compared to healthy controls [33]. Both hypercoagulability and hypocoagulability identified by ROTEM in septic patients were correlated with worse outcomes compared to septic patients with normocoagulability [34].

Besides their role in hemostasis, platelets play important roles in inflammation and immune defense, becoming highly activated in sepsis. Activated platelets expose phospholipids on their membranes contributing to the amplification of coagulation, and release mediators which will activate other circulating platelets. The activated platelets can adhere to the endothelium, bind von Willebrand factor and fibrinogen, trigger the neutrophils to release Neutrophil Extracellular Traps, and form platelet-leucocyte aggregates, bringing an important contribution to microcirculatory damage and intravascular clotting in septic patients [35, 36]. The involvement of platelets in different sepsis-associated organ dysfunctions was already described in literature [36,37,38]. Increased platelet activation was documented in ARDS, in sepsis AKI, and in septic cardiomyopathy in human or experimental studies [39,40,41,42,43,44]. In order to address the platelet contribution to organ dysfunction, antiplatelet strategies were proposed based on positive results from animal studies [45, 46]. However, contradictory results were obtained from human studies [47,48,49,50].

To our best knowledge, there is no previous publication assessing the correlation between platelet counts or hemostatic parameters and early sepsis-associated neurological dysfunction. Our study has some limitations. First, this is a retrospective cohort suitable only for further hypothesis generation. Second, the extended panel of hemostatic tests including plasmatic levels of coagulation factors and inhibitors and viscoelastic tests was not available in the whole study group. The size of the study group having complete hemostatic testing was small and a larger number of septic patients would be necessary to confirm our findings. Third, only onetime point testing was performed, not accounting for the dynamic changes in the coagulation state that are possible and important to follow in these complex patients. Another limitation is the lack of other laboratory methods for the specific quantification of important hemostasis components such as thrombin generation, platelet function tests and enzymatic fibrinolysis.

In early stages of sepsis, a more active coagulation system revealed by faster propagation phase and higher clot firmness and elasticity is correlated with the development of SAE. According to our results, the platelets seem to have an important role, as patients with early SAE presented a higher platelet count, and a higher platelet contribution to clot strength. More research is needed to investigate the role of platelets and the coagulation system in relation with the development of early SAE.