Background
Burns severe enough to require medical attention are observed in nearly 11 million people and represent the fourth most common type of injury globally [
1]. Although most burns are not fatal, more than 300,000 people worldwide do die each year of burn injuries [
1]. The occurrence of coagulopathy in severe burns has been demonstrated in several studies [
2‐
4]. Patients with severe burns are at high risk of inflammation and activation of the coagulation system, and a significant proportion of patients present with coagulopathy unrelated to fluid administration [
5]. The emergence of coagulopathy is an independent predictor of 28-day mortality in patients with severe burns [
2,
3,
5]. However, there is limited evidence regarding effective treatments for coagulopathy in severe-burn patients, and clear treatment recommendations in these patients are lacking [
4].
Several previous studies suggested that antithrombin administration may be effective for treating coagulopathy in severe burns; however, there is currently no robust evidence to support this idea [
6‐
12]. Low antithrombin levels have been identified as an independent predictor of mortality and duration of hospital stay [
13]. Furthermore, a recent study reported therapeutic effects of antithrombin substitution on myocardial dysfunction and inflammation, as well as systemic fluid accumulation, following burns and smoke-inhalation injury in a clinically relevant animal model [
8]. Lavrentieva et al. [
9] conducted a pilot study (
n = 31) to evaluate the efficacy of antithrombin in acute-phase burn-injury patients and found that antithrombin reduced hypercoagulation and improved organ function. However, the effect of antithrombin on mortality in patients with severe burns remains unknown.
We hypothesized that antithrombin could be effective for the treatment of patients with severe burns. The current study aimed to evaluate this hypothesis using data from a large nationwide inpatient database in Japan.
Discussion
In the current study, we analysed data from a Japanese nationwide in-hospital database and found a significant association between antithrombin use and reduction in 28-day mortality in patients with severe burns. This finding was robust with regard to the results obtained by logistic regression and survival analyses. Antithrombin use was also associated with more VFDs in these patients.
The strengths of the current study included its use of a nationwide database and controlling for major factors (e.g. age, size and depth of the burn, and the existence of inhalation injury) that could potentially affect mortality and the extent of haemostatic changes in patients. The extent of haemostatic change is associated with the severity of the burn [
26]; although most patients with mild burns have no systemic coagulation changes, patients with severe burns frequently develop coagulopathy [
26]. We therefore only included patients with severe burns (i.e. burn index >10) [
15] in the current study. In addition to burn size, full-thickness burns and inhalation trauma are also associated with the occurrence and severity of systemic coagulopathy [
5,
32]. The baseline patient characteristics in the unmatched groups suggested that antithrombin use was higher in patients with severe burns (e.g. those with higher mortality and higher burn index and those requiring mechanical ventilation, catecholamines, and other treatments). However, we ensured the comparabilities of the two groups by propensity score matching, which provides a sound method for constructing a randomized experiment-like situation by comparing groups with similar observed characteristics, without specifying the relationships between confounders and outcomes [
24]. We included factors that could potentially affect mortality and the extent of haemostatic changes in patients with severe burns when estimating propensity scores, such as age, burn-surface area and depth (i.e. burn index), comorbidities, and mechanical ventilation status [
15,
19,
28]. After one-to-one propensity score matching, the two groups (i.e. antithrombin vs. control) were well balanced in most of the measured variables. Our results suggested that patients with severe burns who received antithrombin were less likely to die and had more VFDs than those who did not receive antithrombin. There was no significant difference in the incidence of post-admission complications between two groups. The current results were compatible with those of previous sepsis studies, which found that antithrombin did not increase bleeding risk, at least at the dose used in Japan (i.e. 1500 U/day) [
33,
34].
Several previous studies suggested that severe thermal injury is associated with early activation of the coagulation cascade and the emergence of coagulopathy, which worsens the outcome of burns patients [
2‐
5]. Interestingly, several similar patterns of procoagulant and antifibrinolytic changes, as well as natural anticoagulant system impairments, were witnessed in both early-stage severe-burn patients and sepsis patients [
2,
3,
9,
35]. The guidelines of the Japanese Society of Thrombosis Haemostasis [
36] and the Japanese Society of Intensive Care Medicine [
37] both recommend the use of supplemental antithrombin in patients with sepsis-associated coagulopathy, and anticoagulant therapy, including antithrombin, is thus commonly used in clinical practice in Japan [
34]. Although no large randomized trials have suggested the effectiveness of antithrombin use, recent nationwide database studies suggested an association between lower in-hospital mortality and supplemental antithrombin administration among patients with pneumonia and abdominal sepsis [
16,
29].
Several treatment guidelines for burns, including Japanese guidelines [
38], do not address or recommend the use of anticoagulation treatments for severe-burn patients, because of a lack of scientific evidence [
39]. European guidelines for bleeding care in trauma management suggest that antithrombin is not recommended in acutely bleeding trauma patients because of the increased risk of bleeding events and its failure to reduce overall mortality [
40]. Although numerous studies have explored coagulopathy in sepsis patients, few have evaluated coagulopathy in severe-burn patients in any detail [
4]. Moreover, clear definitions or diagnostic criteria for coagulopathy after burns are currently absent. A lack of simple and easy-to-interpret diagnostic tests may thus disguise the true incidence of coagulopathy in patients with severe burns [
4]. Antithrombin deficiency is common after burns and is related to total burn-surface area and inhalation injury, increased mortality, and longer hospital stay [
13]. However, treatments for coagulopathy after severe burns may be less popular than treatments for sepsis-associated coagulopathy. Indeed, the current nationwide study, using data from a country where antithrombin is commonly used as anticoagulant therapy in clinical practice [
34], found that antithrombin was only administered to 152 of 3223 eligible severe-burn patients. The current results may thus support further prospective studies in this area.
A previous study suggested that early-onset coagulopathy was associated with prolonged mechanical ventilation [
32], while the current results indicated that antithrombin use was associated with more VFDs. This may be because antithrombin acts as an anticoagulant with anti-inflammatory properties. A previous observational study [
41] showed that pulmonary coagulopathy seemed intrinsic to burn injuries and inhalation trauma, and concluded that patients with burn injuries and inhalation trauma requiring mechanical ventilation showed a distinct and sustained procoagulant and antifibrinolytic shift in the pulmonary compartment, indicating that pulmonary coagulopathy could be an important therapeutic target in these patients. Several recent experimental studies suggested that antithrombin attenuates myocardial dysfunction and pulmonary vascular leakage in patients with severe burns and inhalation injury [
8,
42]. Rehberg et al. [
42] showed an interaction between antithrombin and neutrophils in vivo and demonstrated its pathophysiological role in vascular leakage and the therapeutic potential of antithrombin in a sheep model. Rehberg et al. [
8] also demonstrated therapeutic effects of antithrombin in terms of myocardial dysfunction and inflammation, as well as systemic fluid accumulation following burns and smoke-inhalation injury in a sheep model. Based on the current and previous studies [
8,
41,
42], supplemental antithrombin treatment may represent a valuable therapeutic approach for cardiovascular dysfunction and inflammation after burn and smoke-inhalation injury. However, further studies are required to confirm these results.
This study had some limitations. First, the current study was retrospective and observational in nature, without randomization. Even though propensity score matching was used to adjust for differences in baseline characteristics and disease severity, including factors that could potentially affect mortality and the extent of haemostatic change in patients with severe burns, there may still have been bias in the form of unmeasured confounders. Possible confounding parameters include laboratory-based coagulation tests, which indicate the level of coagulopathy (e.g. international normalized ratio, activated partial thromboplastin time, antithrombin), systemic inflammatory syndrome response, disseminated intravascular coagulation scoring system, blood gas analysis, serum lactate level, fluid balance per day, haemodilution caused by large amounts of resuscitation fluids together with hypothermia, and type of hospital admission (e.g. general burn ward or advanced burn unit). Unfortunately, these data were not available from the DPC database. However, previous studies suggested that the severity of coagulopathy was correlated with the extent and depth of the burn and the existence of inhalation injury [
4,
26], and we were able to match these factors using propensity scoring, when comparing the outcomes between the antithrombin and control groups. Second, this study only evaluated the association between early-phase antithrombin use and outcomes and did not evaluate the effect of later-phase antithrombin use (e.g. antithrombin use after sepsis-associated disseminated intravascular coagulation due to burn-wound infection). Third, we could not identify the cause of death or long-term outcome from the current DPC database and therefore could not speculate on the factors responsible for the improved 28-day mortality in the antithrombin group compared with the control group. Fourth, we did not determine whether the replacement of coagulation factors using fresh-frozen plasma was superior to antithrombin provision, though this research question should be evaluated in future studies.
Authors’ contributions
TT and HY had full access to all the study data and take responsibility for the integrity of the data and the accuracy of the data analysis. All authors contributed to study concept and design. TT, HM, KF, and HY helped in acquisition and analysis of data. All authors interpreted data. TT drafted the manuscript. All authors critically revised the manuscript for important intellectual content. All authors read and approved the final manuscript.