Plasma protein C levels are directly associated with better outcomes in patients with severe burns

burns 45 (2019) 1659 –1672

Available online at www.sciencedirect.com

ScienceDirect journal homepage: www.elsevier.com/locate/burns

Plasma protein C levels are directly associated with better outcomes in patients with severe burns $

Thomas Charles Lang a,c, * , Ruilong Zhao a, Albert Kim b, Aruna Wijewardena b, John Vandervord b , Rachel McGrath b , Siobhan Fitzpatrick b , Gregory Fulcher b, Christopher John Jackson a a Sutton Laboratories Level 10, The Kolling Institute, The University of Sydney, Northern Clinical School, Royal North Shore Hospital, Reserve Rd, St. Leonards, 2065, NSW, Australia b Royal North Shore Hospital, Reserve Rd St., Leonards, 2065, NSW, Australia c Department of Anaesthesia, Prince of Wales and Sydney Children’s Hospitals, Barker St, Randwick, 2031, NSW, Australia

article info

abstract

Article history:

Protein C circulates in human plasma to regulate inflammation and coagulation. It has

Accepted 1 May 2019

shown a crucial role in wound healing in animals, and low plasma levels predict the presence of a wound in diabetic patients. However, no detailed study has measured protein C levels in patients with severe burns over the course of a hospital admission. A severe burn is associated with dysfunction of inflammation and coagulation as well as a significant risk of

Keywords:

morbidity and mortality. The current methods of burn assessment have shortcomings in

Protein C

reliability and have limited prognostic value. The discovery of a biomarker that estimates

Severe burns

burn severity and predicts clinical events with greater accuracy than current methods may

Surgery

improve management, resource allocation and patient counseling. This is the first study to

Clinical outcomes

assess the potential role of protein C as a biomarker of burn severity.

Intensive care

We measured the plasma protein C levels of 86 patients immediately following a severe

Fluid resuscitation

burn, then every three days over the first three weeks of a hospital admission. We also analysed the relationships between burn characteristics, blood test results including plasma protein C levels and clinical events. We used a primary composite outcome of increased support utilisation defined as: a mean intravenous fluid administration volume of five litres or more per day over the first 72 h of admission, a length of stay in the intensive care unit of more than four days, or greater than four surgical procedures during admission. The hypothesis was that low protein C levels would be negatively associated with increased support utilisation. At presentation to hospital after a severe burn, the mean plasma protein C level was 76  20% with a range of 34–130% compared to the normal range of 70–180%. The initial low can be plausibly explained by impaired synthesis, increased degradation and excessive consumption of protein C following a burn. Levels increased gradually over six days then remained at a steady-state until the end of the inpatient study period, day 21. A multivariable regression model (Nagelkerke’s R2 = 0.83) showed that the plasma protein C level on admission contributed the most to the ability of the model to predict increased support utilisation

$

The Department or Institution to which this work should be attributed is The Kolling Institute of The University of Sydney, associated with Royal North Shore Hospital. * Corresponding author at: Department of Anaesthesia, Prince of Wales and Sydney Children’s Hospitals, Barker St, Randwick 2031, NSW Australia. E-mail address: [email protected] (T.C. Lang). https://doi.org/10.1016/j.burns.2019.05.001 0305-4179/© 2019 Elsevier Ltd and ISBI. All rights reserved.

1660

burns 45 (2019) 1659 –1672

(OR = 0.825 (95% CI = 0.698-0.977), P = 0.025), followed by burn size (OR = 1.252 (95% CI = 1.025– 1.530), P = 0.027), burn depth (partial thickness was used as the reference, full thickness OR = 80.499 (1.569–4129.248), P = 0.029), and neutrophil count on admission (OR = 1.532 (95% CI = 0.950–2.473), P = 0.08). Together, these four variables predicted increased support utilisation with 93.2% accuracy, 83.3% sensitivity and 97.6% specificity. However if protein C values were disregarded, only 49.5% of the variance was explained, with 82% accuracy, 63% sensitivity and 91.5% specificity. Thus, protein C may be a useful biomarker of burn severity and study replication will enable validation of these novel findings. © 2019 Elsevier Ltd and ISBI. All rights reserved.

1.

Background

A severe burn is associated with dysfunction of both inflammation and coagulation and a significant risk of morbidity and mortality. Uncontrolled inflammation following burns leads to the systemic inflammatory response syndrome [1] and a paradoxically increased risk of infection and sepsis [2]. Disseminated intravascular coagulation, microvascular thrombosis, haemoconcentration [3], ineffective fibrin aggregation, platelet dysfunction [4] and aggressive fluid resuscitation primarily with crystalloid solution [5] all contribute to widespread disruption to coagulation processes and poor tissue perfusion. Both inflammatory and coagulation dysfunction pathways converge to produce, in many cases, multi-organ dysfunction syndrome [6] These complications benefit from intensive and coordinated treatment to prevent morbidity and mortality. Adjusted inpatient mortality incidence is approximately 4%. [7] This is similar to the inpatient mortality incidence among patients with acute myocardial infarction [8] or pulmonary embolus [9]. The adjusted all-cause mortality rate ratio is 1.8 times higher among adolescent, adult and middleaged patients with burns compared to an index hospital population without burns [10]. Among older patients, longterm mortality attributable to burn injury is 29% [11]. Morbidity is prevalent but more difficult to quantify given the heterogenous nature of the condition and the variety of methods used to report outcomes after a burn [12–14]. Given this risk of mortality and morbidity, prognostic models are used in the care of patients with burns. These models are conventionally derived from regression analyses of clinical data [15]. Such data includes findings from the clinical examination like burn size and burn depth. Prognostic models are often referred to as burn severity scores. However, many scores are not properly validated, do not provide odds ratios and cannot quantify the risk of a given outcome [16]. Therefore, other data should be used in conjunction with the clinical examination to accurately predict clinical events. The discovery of a biomarker, which provides clinicians with an adjunct to the clinical examination in order to estimate burn severity, quantify risk and predict clinical events with greater accuracy than current methods, may improve management decisions, resource allocation and patient counseling. Indeed, studies have shown an association between clinical events following a sever burn and the results of blood tests. For example, procalcitonin predicts

sepsis after a burn, [17,18] and inflammatory cytokine levels such as IL-8, [19,20], IL-4, GM-CSF, MCP-1 are associated with mortality [20] as are enzymes TIMP-1 [21] and gelsolin [22]. Protein C is a serine protease synthesised largely by the liver [23], as well as vascular endothelial cells and epidermal keratinocytes [24], which circulates in plasma as a zymogen of activated protein C (APC) [23]. It has a molecular weight of 62 kD and consists of 419 amino acids: three major moieties comprise the light chain, which connects to the catalytic heavy chain by a disulphide bridge [23]. The protein C system is best known for its function of regulating coagulation [25]. In addition, APC exerts potent anti-inflammatory and cytoprotective functions [23]. Protein C circulates in normal plasma at an average concentration of 4 mg/mL [26]. The plasma concentration of APC is 2000 times lower than the concentration of protein C [23] and measurement is not feasible in clinical practice as the half-life of APC is only 15 min [27]. Protein C levels are far more stable and are easily measured by analysing plasma with an assay which determines functional protein C levels, and reporting protein C levels as a percentage compared to normal levels [26]. The current World Health Organisation (WHO) standardised units for reporting protein C are IU/mL [28] although it is also reported with units of mg/mL [26], % [29] and U/dL [30]. We have shown previously that protein C levels are an independent predictor of the presence of a wound in age, sex and type of diabetes-matched subjects, to a greater extent than HbA1c% or C-reactive protein (CRP) levels [31]. However, diabetic wounds are typically chronic and there is limited data about plasma protein C levels in patients with acute wound states, such as a burn injury. Protein C levels in sheep with experimentally induced burns decreased over the first 24 h after injury [32]. A small study of five patients with burns of 20–70% total body surface area (TBSA) found that protein C levels ranged from 26 to 89% of normal [33]. Another more recent small study of nine patients with burns of 25–95% TBSA found that protein C levels taken at hospital admission ranged from 30 to 95% of normal [29]. Interestingly, this indepth study of dozens of clotting factors and proteins showed that protein C was the only component of the coagulation system that was significantly different between survivors and non-survivors at time of admission (82  9% vs. 56  18%, P = 0.04) [29]. There is physiological congruity between the role of protein C and the pathophysiology of burns, and there is the potential for improvement of methods estimation of severity of injury and prognostication in patients with burns, and

burns 45 (2019) 1659 –1672

there is a growing body of evidence showing prognostic utility of several blood markers of inflammation in burns. Therefore, we measured protein C levels of patients with severe burns on admission and serially throughout their hospital stay and then investigated the relationship between patient characteristics, burn characteristics, several blood test results including protein C levels, and clinical outcomes, in order to test our hypothesis. Our hypothesis was that protein C levels would be negatively associated with support utilisation in patients with severe burns. Other aims of the study included the quantification of protein C levels over time in patients with burns, as well as comparison of protein C levels between groups stratified by burn size, burn depth and survival, as well as the establishment of predictors of outcome following a severe burn based on a multivariable regression model of the collected data.

2.

Methods

This was a single-centre prospective observational cohort study undertaken in Royal North Shore Hospital, a large teaching hospital in New South Wales, Australia, which also functions as a Level 1 Trauma Centre and Burns Referral centre. Participating departments included the Severe Burns Unit (SBU), the Intensive Care Unit (ICU) and the Emergency Department (ED). The SBU provides ward-level care for patients with severe burns admitted under the Burns surgical team. It is staffed with specialist burns staff, who also see patients in the SBU outpatient clinic for follow up. The study was designed and completed in accordance with the Strengthening Reporting of Observational Studies in Epidemiology (STROBE) statement [34] and approved by the Northern Sydney Local Health District Human Research Ethics Committee. Patients with severe burns were admitted to hospital and after providing written informed consent were enrolled as participants according to the following inclusion and exclusion criteria: participants were above eighteen years of age and had burns between 10–80% TBSA involved, with some area of the burn as partial thickness. These burn sizes were selected on the rationale that burns below 10% TBSA would not lead to an appreciable change in protein C levels, while the cut-off of 80% was decided by our ethics committee as a suitable upper limit for patient inclusion. Burn depth was determined clinically by consensus among senior burns surgeons at the time of presentation and was defined as the most prevalent depth of burn. If there was conjecture about burn size or depth, or difficulty with clinical examination in the ED, the burn size and depth were clarified when the patient was under general anaesthesia prior to excision and grafting. This practice is part of usual care at this centre. Patients were not eligible to be enrolled if they were under eighteen years of age, pregnant or lactating, had clinically significant clotting or bleeding disorders (including any pre-existing condition requiring anticoagulation with warfarin or other oral anticoagulants) or had active local or systemic infection. The age cut-off was determined by the ethics committee while pregnancy, infection and clotting disorders can all affect baseline protein C levels

1661

therefore patients were excluded if any were present, in order to maximize the likelihood of including patients with normal baseline protein C levels. Participants were deemed to have severe burns by definition of transfer criteria to this facility in accordance with the New South Wales Burns Transfer Guidelines [35]. Ethics committee approval was gained for the recruitment of 86 patients into this study and was adequately powered for data analysis of approximately 60 patients, based on results from a similar study, using an a prioi analysis [31]. Selection bias was minimised by the prospective nature of enrolment and strict adherence to inclusion and exclusion criteria, and loss-to-follow up bias was minimised by active retention and regular follow up in our SBU outpatient clinic. Potential confounders were minimised by recording all exposures during the patient’s admission and the cohort underwent stratification and regression. Patients were assessed immediately on arrival to the ED by senior burns surgeons as part of usual care, which included visual assessment of the burn with formal quantification of burn size and depth. All patients received usual care throughout and after the inpatient study period of 21 days. Blood samples were collected on the day of admission (day 0) and subsequently every three days for three weeks, until day 21 or discharge. Time of hospital presentation, rather than time of injury, was selected as time point zero for several reasons. Firstly, we lacked accurate data for all patients about the time of injury. Secondly, the risk of data collection and data entry error was felt to be unacceptably high if time of injury was selected as time point zero, because all electronic medical record data used time of presentation as time point zero. Thirdly, the estimated time of injury occurred within 24 h to presentation to hospital for all patients. Based on the understanding of human protein C production and consumption, this was felt to be an acceptable time-window before the first blood sample was obtained. Blood samples were taken third-daily, rather than another frequency, based on the rationale that more frequent collection would be unlikely to yield more valuable data for the purposes of the study, whilst also being unlikely to yield clinically valuable data and unacceptably increasing patient discomfort. Blood samples were also taken at followup at 12 months after discharge. Blood samples were obtained and analysed for the total amount of functional protein C by trained pathologists in the Royal North Shore Hospital Pacific Laboratory Medicine Services using a Diagnostica Stago STA-R Evolution (France) coagulation analyser (Series Number CA81044175). Blood samples were included for analysis if they were collected within 24 h of each designated time point. Clinical data were also collected as part of routine care and included demographic details, intravenous fluid, details of surgery undertaken, length of stay (LOS) within the hospital and within the ICU and several blood markers of inflammation and organ function. 1 1 Data was analysed with IBM SPSS Statistics v20 and 1 GraphPad Prism v7. A statistician not involved in data collection independently oversaw analysis of the data. Statistical tests were undertaken to assess the relationships between patient characteristics, laboratory data and clinical events.

1662

burns 45 (2019) 1659 –1672

2.1. Primary outcome: increased support utilisation throughout admission We used a binary composite outcome for this study, termed increased support utilisation. This outcome identified patients who received large amounts of intravenous fluid, and/or had a prolonged LOS in the ICU, and/or had numerous surgical procedures during their admission. Each of these components of care is considered standard treatment but each carries risks, which contribute to patient morbidity and therefore excess exposure to these treatments comprises a complicated admission utilising increased support. It is important to identify these clinical events in order to carefully consider exactly how best to select the best treatment for these patients.

2.1.1.

Volume of intravenous fluid administered

In patients with severe burns, complications of high-volume fluid resuscitation are well described and include pulmonary oedema [36], intraabdominal hypertension (IAH) and abdominal compartment syndrome (ACS) [37–42]. Fluid infusion regimes such as the Parkland formula provide high volumes of fluid in the first 8 h, with the volume administered over the subsequent 16 h guided by urine output [43]. Most burns centres, including ours, utilise the Parkland protocol [44,45]. However, the majority do not adhere to the original protocol and subsequently, patients are administered too much fluid [46–48]. There is clear evidence of a relationship between the volume of intravenous fluid administered, particularly crystalloid solutions, and the incidence of adverse events. A lack of validated protocols for fluid administration in patients with severe burns beyond the first 24 h after injury may contribute to this phenomenon. Klein et al. found a total volume of 250 mL/kg of crystalloid fluid led to patients having a higher risk of death, acute respiratory distress syndrome, pneumonia, and other complications [49]. Ivy et al. found a threshold of 250 mL/kg of fluid led to ACS at 48 h after burn [50]. Oda et al. found a fluid threshold of 300 mL/kg at any time point in admission to produce IAH, which in their cohort was approximately 21 L of intravenous fluid administered throughout admission. In that study, patients received intravenous fluid infusions for at least 4 days which is equivalent to 5.25 L per day for four days in a 70 kg patient [51]. In a randomised trial, surgical patients who were administered higher volumes of intravenous fluids had higher rates of complications compared to patients who were administered restricted volumes of fluid and this was most marked for patients who gained >5.51 kg of body weight with the amount of weight gain in kg closely related to the volume of intravenous fluid administered in L, with the highest risks of complications occurring in the first 72 h [52]. Given that the usual duration of fluid resuscitation in patients with burns is at least three days [51], the average ideal body weight in a patient with burns is approximately 70 kg [53], and the total volume of intravenous fluid which is associated with morbidity is approximately 250 mL/kg, the approximate time point at which a high-risk volume of fluid of is administered is three days. Therefore, we instituted part of our primary outcome to include a mean intravenous fluid administration volume of 5 L or more per day over the first 72 h of admission.

2.1.2.

ICU LOS

Mortality rates remain steady [7] therefore studies of predictors of patient outcomes focus on morbidity. Measures of morbidity are not equal. Some are easy to measure, like hospital LOS. However, hospital LOS alone does not accurately capture the degree of support during an admission and does not impact the cost of a hospital admission, whereas ICU LOS does [54]. LOS in the ICU provides a clearer picture of a patient who utilises significant support during their admission and it also predicts mortality after discharge [55,56]. More recent studies of patients with burns have used ICU LOS as their primary outcome [57] or as part of a composite primary outcome [58–61]. For example, one study of 250 patients with a mean TBSA of 32.2% showed that patients spent a mean duration in ICU of 0.23 days per 1% TBSA [54]. ICU LOS is important to determine the amount of support a patient needs, the cost of admission, and predicting mortality after discharge. Furthermore, longer stays in ICU are associated with adverse outcomes. In 821 adult patients with respiratory failure or sepsis without known underlying cognitive deficit and a median ICU LOS of five days, 34% of patients exhibited signs of a moderate traumatic brain injury and 24% exhibited signs of Alzheimer’s disease at one year follow up [62]. In 34,696 elderly ICU patients with a mean LOS in ICU of 3.4 days, each day beyond seven days in ICU increased the odds of death at one year by 104% [63]. A systematic review of sequential organ failure assessment (SOFA) models in patients in ICU showed that severity of illness measured over five days correlated strongly with mortality [64]. Thus we included patients in the outcome of increased support if they had an ICU LOS of five or more days.

2.1.3.

Number of surgical procedures undertaken

Practices vary between centres regarding the timing and amount of surgery undertaken [65]. In our centre, almost all patients with severe burns undergo excision and grafting. Graft loss is a complication of excision and grafting following a burn. Among patients with graft loss, re-operation is common [66]. Most of our procedures are performed under general anaesthesia. It is known that perioperative complications increase with the number of surgeries during admission [67–69] such as pathological scarring [70] and general anaesthesia associated post-operative cognitive dysfunction syndrome [71–73]. There is limited data to suggest a threshold number of surgeries that predisposes someone to increased risk of adverse outcomes and further studies are required. So in order to identify patients who may have graft failure and who may potentially be at risk of complications from multiple general anaesthetics, we defined a group of patients who had a higher than expected number of surgeries. The mean number of surgeries each patient underwent in our study was four, so we included patients in this outcome who underwent more than the mean of four surgical procedures. Thus, we instituted a composite binary classification to identify patients who used increased support throughout admission. This included patients with any of the following: a mean intravenous fluid administration of 5 L or more per day over the first 72 h of admission, a LOS in the ICU of five or more days, or five or more surgeries during the admission.

1663

burns 45 (2019) 1659 –1672

2.2.

Table 2 – Clinical overview of patient treatment throughout their admission.

Statistics

Descriptive statistics are expressed with the mean followed by standard deviation, or by n followed by %, unless otherwise stated. Statistical tests used are described when relevant.

3.

Results

3.1.

Patient and burn characteristics

All 86 enrolled patients who met inclusion and exclusion criteria were included in the prospective analysis. Patient and burn characteristics are shown in Table 1. One quarter of patients were female and the cohort had a mean age of 44 years. The most common type of burn was thermal (90%) and the mean burn size was 21% of TBSA. Two-thirds of patients had primarily partial thickness burns (66%) and almost a third of patients had full thickness burns (32%). Table 3 summarises the results of commonly performed blood tests.

3.2.

All patients Intubated, n (%) Yes No Admitted to, n (%) SBU ICU LOS (days), mean  SD Total ICU Duration of intubation in days, mean  SD Intravenous fluid administration (L), mean  SD Day 0 Day 1 Day 2 Surgical management, n (%) Yes No Number of surgeries, mean  SD Mortality, n (%)

23 (27) 63 (73) 49 (57) 37 (43) 25  31 6  14 88 3.64  3.09 4.14  2.94 3.23  1.93 76 (88) 10 (12) 43 3 (3)

Patient treatment

Table 2 summarises some key features of patient treatment through the phases of initial resuscitation, intensive care management, and surgical management. Evidence of an airway burn was found in 25 patients (29%) on presentation and endotracheal intubation was performed in 23 of those patients (27%). Over half of the patients (57%) were admitted to the ICU and the remainder were admitted to the SBU. Of patients admitted to the ICU, the mean LOS was six days. The overall mean hospital LOS for the entire cohort was 25 days. Among patients who were intubated, the mean duration of intubation and ventilation was 8 days. The mean volume of intravenous fluid provided to patients during resuscitation and their first day in hospital was 3.64 L, with a mean of 4.14 L then 3.23 L on subsequent days. Surgical intervention was

Table 1 – Clinical overview of patient and burn characteristics on admission. All patients Total patients, n Female, n (%) Male, n (%) Age (years), mean  SD Mechanism of injury, n (%) Thermal — flame Thermal — scald Thermal — other Chemical Electrical Friction Burn size (% TBSA), mean  SD Primary depth of burn, n (%) Full thickness Partial deep Partial superficial Superficial

86 22 (26) 64 (74) 44  19 65 (76) 9 (10) 3 (3) 7 (8) 1 (1) 1 (1) 21  13 27 (31) 31 (36) 26 (30) 1 (1)

Table 3 – Baseline laboratory results for blood tests on presentation to hospital. All patients 9

Neutrophils 10 Mean  SD Normal range 2–8  109 Lymphocytes 109 Mean  SD Normal range 1–4  109 Platelets 109 Mean  SD Normal range 150–400  109 Albumin (g/L) Mean  SD Normal range 35–52 g/L Creatinine (mmol/L) Mean  SD Normal 45–90 umol/L C-reactive protein (mg/L) Mean  SD Normal <5 mg/L INR (International normalized ratio) Mean  SD APTT seconds Mean  SD Normal range 24–36 s Lactate (mmol/L) Mean  SD Normal 0.5–1.6 mmol/L

10.7  6.1

2.1  1.1

240  84

37  6

80  31

39  52

1.1  0.2 28.9  4.1

2.0  1.4

undertaken in 76 patients (88%) and of these patients the mean number of surgeries was four. Three patients did not survive admission.

3.3.

Protein C levels on admission

On admission, the mean protein C level for all patients was 76  21% with a range of 34–130% compared to the normal

1664

burns 45 (2019) 1659 –1672

range of 70–180% [74]. Values were normally distributed. Twenty-one of the 59 patients (35%) who provided admission blood samples had lower than normal levels of protein C (<70%). These patients had a mean protein C level of 56  12% on admission. Eleven patients had full thickness burns, ten had burns of partial thickness, and there was a mean burn size of 27  18% TBSA. The mean daily intravenous fluid administration volume over the first 72 h of admission was 4.3  2.3 L/ day, and the mean number of surgeries was 6  5 surgeries. These patients had a mean ICU LOS of 16  24 days and a mean total hospital LOS of 41  44 days. All three non-survivors had admission protein C levels <70% on admission.

3.4.

Protein C levels throughout admission

Fig. 1 and Table 4 outline plasma protein C levels in this cohort throughout the 3-week study period of their admission. Regarding the change in protein C values among the entire cohort over the 21 day study period, an ordinary one-way ANOVA analysis (which was significant, P = 0.0003) followed by Dunnett’s multiple comparisons showed a statistically significant increase in mean protein C level compared to admission by day 6 and this difference remained significant until day 21.

Fig. 1 – Plasma protein C levels in patients with severe burns. In a cohort of 86 patients with severe burns with a mean TBSA % of 2113, plasma protein C levels were measured every three days over 21 days of a hospital admission. Asterixes represent a statistically significant increase in mean protein C level from admission. ****P <0.001. Bars show SEM.

We explored the relationship between plasma protein C and burn depth and size. Protein C levels were lower in full thickness burns than partial thickness burns on admission (64  18% vs. 82  19%, P = 0.0008) and this difference remained statistically significant up to and including day 6. Patients with larger burns (defined as TBSA  30% [37,75]) had lower protein C levels than patients with smaller burns (TBSA < 30%) on admission (55  17% vs. 80  19%, P = 0.0003), and this difference remained statistically significant up to and including day 12. These results are summarised in Fig. 2a and b. Table 5 and Fig. 2d outline the protein C levels of three patients (3%) who died during their hospital admission. Two of these patients had protein C levels taken on presentation to the hospital (49% and 47% respectively). Their protein C levels remained low throughout admission. Statistical tests were not undertaken due to the low number of values for each time point but it is clear that these patients’ protein C levels were lower than those of survivors, and remained very low throughout admission.

3.5. The association between protein C levels on admission and increased support use Patients requiring increased support throughout admission had significantly lower levels of protein C on admission than those who did not (55  17% vs. 80  18%, P < 0.0001, Fig. 2c) and this difference remained statistically significant up to and including day 12. Table 6 shows the patient characteristics of those who underwent increased support, and those who had a standard admission. There was a statistically significant difference between these two groups with regards to patient and burn characteristics, pathology results, and treatment details; mean burn size and depth were significantly different between the two groups, as were the proportion of patients who were intubated and sent immediately to ICU from the ED (each group P < 0.0001). Table 7 shows blood test results on admission which showed a significant difference between patients who went on to use increased support, and patients who did not, were as follows: protein C level (P < 0.0001); neutrophil count (P < 0.0001); creatinine (P = 0.012) and lactate (P = 0.0231). The following blood test results were not significantly different between the two groups on presentation to hospital: CRP (P = 0.0938); platelet count (P = 0.3997); INR (P = 0.3842); lymphocyte count (P = 0.4899); albumin level (P = 0.8054) and APTT (P = 0.7602).

Table 4 – Protein C levels in patients with severe burns throughout hospital admission. Day

n

Mean  SD (%)

Range (%) Normal = 70–180%

Adjusted P value (comparison to day 0 value)

0 3 6 9 12 15 18 21

59 65 69 56 47 34 30 23

76  20 89  25 106  30 116  33 117  33 116  39 121  36 114  36

34–130 45–135 41–183 40–185 52–200 32–185 31–193 31–170

– 0.1197 <0.0001 <0.0001 <0.0001 <0.0001 <0.0001 <0.0001

burns 45 (2019) 1659 –1672

1665

Fig. 2 – The cohort was stratified according to: burn depth; burn size; the need for increased support throughout hospital admission; and mortality. The plasma protein C level is shown among each of these groups. Increased support was defined as any of a mean intravenous fluid administration volume of >5L/day over the first 72h of hospital admission, five or more days spent in ICU, or five or more surgical procedures undertaken during admission. Groups were analysed for statistically significant differences in protein C levels at each time point. Asterixes represent a statistically signifiicant differece in mean between groups at each time point. ****P <0.001; ***P <0.01; **P <0.02; and *P <0.05. Bars show SEM. (a) Plasma protein C levels in patients with full thickness burns compared to patients with partial thickness burns. Plasma protein C levels were lower in patients with full thickness burns (black line) compared to patients with partial thickness burns (grey line) until day 9. (b) Plasma protein C levels in patients with large burns compared to patients with small burns. Plasma protein C levels were lower in patients with large (30% TBSA) burns (black line) compared to patients with small (<30% TBSA) burns (grey line) until day 12. (c) Plasma protein C levels in patients with severe burns requiring increased support compared to those who did not. Plasma protein C levels were lower in patients who had increased support throughout admission (black line) compared to patients who did not require increased support (grey line) until day 12. (d) Plasma protein C levels in patients with severe burns who died during admission compared to patients who survived. Plasma protein C levels appeared lower in patients who died during their hospital admission than those who survived. Statistical tests were not performed due to the small sample size of patients who died.

Table 8 shows the clinical events that occurred throughout admission among the two groups, and unsurprisingly, because the events that occurred during the admission defined each group, there were multiple statistically significant differences in the occurrence of events such as amount of fluid administered, number of surgeries, LOS and LOS in ICU, and mortality incidence. Regression analysis (Table 9) was performed to ascertain the extent to which patient characteristics, burn characteristics and blood test results on admission determined a patient’s outcome after adjustment for the influence of other variables. Out of the total 86 patients, 57 patients could be included in the stepwise analysis due to missing data, and of these, 19 were identified as using increased support. We identified, with a comprehensive and systematic method

exploring plausible factors which may influence the outcome, the following variables that significantly increased the risk of patients with increased support: burn size (n = 86), burn depth (n = 86), presence of inhalational injury (n = 86), day 0 protein C (n = 57), day 0 neutrophil count (n = 88), day 0 lactate (n = 55) and day 0 creatinine (n = 87). Table 10 shows the final four variables on admission that contributed most to influencing increased support throughout admission based on multiple imputations of several regression model combinations. These variables are: protein C level on admission, burn size, burn depth, and neutrophil count on admission. Of these, protein C levels contributed the greatest log likelihood change for the regression modelling. This model’s performance was statistically significant, P < 0.0005. It was then evaluated by several methods. Firstly, Nagelkerke’s

1666

burns 45 (2019) 1659 –1672

Table 5 – Protein C levels in patients with severe burns who died during hospital admission. Day

n

Mean  SD (%)

Range (%)

0 3 6 9 12 15 18 21

2 3 2 2 0 2 1 2

48  1 56  10 69  9 68  40 – 81  18 81  0 71  25

47–49 45–62 62–75 40–96 – 68-94 81 53–88

R [2] (0.83) showed that the model explained 83.1% of the variance in outcome, and correctly classified 93.2% of cases of increased support. Sensitivity was 83.3% and specificity was 97.6%. Secondly, the Hosmer-Lemeshow goodness-of-fit test for calibration was appropriately not significant, P = 0.907. Thirdly, the area under the receiver-operator characteristic (ROC) curve (AUC) for discrimination was 0.974 (95% CI, 0.940– 1.000), P < 0.0005; and the optimism-corrected ROC AUC was 0.937. These are outstanding levels of discrimination [76]. Internal validation was achieved by calculating the optimismcorrected ROC AUC from 200 bootstrapped samples. Further consideration of the Youden index from protein C’s ROC curve suggested that an optimal protein C cut-off level of 70% leads to the ideal sensitivity and specificity values for the model in predicting the risk of increased support. This coincidentally is the lower limit in normal in adults provided by our laboratory, when testing patients with venous thrombosis for inherited protein C deficiency. When considering the performance of the multivariate regression model in the absence of protein C, the model’s performance deteriorated. It remained statistically significant (P < 0.0005) however the goodness-of-fit test for

Table 7 – Baseline blood test results on admission. Increased support

P value

85  16

57  17

<0.0001

8.9  4

14.7  7.9

<0.0001

2.1  1.2

1.9  0.9

0.4899

235  71

252  108

0.3997

74  15

92  45

0.012

37  7

37  6

0.8054

52  70

24  32

0.0938

1.1  0.1

1.1  0.2

0.3842

28.8  2.7

29  6.1

0.7602

1.6  0.7

2.5  1.9

0.0231

Standard admission Protein C (%) Mean  SD Normal range 70–180 Neutrophils 109 Mean  SD Normal range 2–8 Lymphocytes 109 Mean  SD Normal range 1–4 Platelets 109 Mean  SD Normal range 150–400 Creatinine (mmol/L) Mean  SD Normal range 45–90 Albumin (g/L) Mean  SD Normal range 35–52 C-reactive protein (mg/L) Mean  SD Normal <5 INR Mean  SD APTT (seconds) Mean  SD Normal range 24–36 Lactate (mmol/L) Mean  SD Normal range 0.5–1.6

P values formatted to bold font are all values <0.05.

Table 6 – Clinical overview of patients on admission. P value represents comparison of increased support vs. standard admission. P value

Standard admission

Increased support

Total patients, n 59 14 (24) Female, n (% of group) 45 (76) Male, n (% of group) 43  19 Age (years), mean  SD Mechanism of injury, n (%) 43 (72) Thermal (flame) 9 (15) Thermal (scald) 2 (3) Thermal (other) 5 (8) Chemical 0 (0) Electrical 1 (2) Friction Burn size (% TBSA), 17  8 mean  SD Primary depth of burn 10 (17) Full thickness Partial deep 25 (42) 24 (40) Partial superficial 1 (1) Superficial

27 8 (30) 19 (70) 47  17

– 0.6003

24 (86) 0 (0) 1 (4) 2 (7) 1 (4) 0 (0) 31  19

0.1965

18 (64) 8 (26) 2 (7) 0 (0)

<0.0001

P values formatted to bold font are all values <0.05.

0.4233

<0.0001

calibration trended towards significance (P = 0.115), Nagelkerke’s R [2] reduced to 0.495, the classification accuracy reduced to 82.6%, the sensitivity reduced to 63.0% and the specificity reduced to 91.5%.

4.

Discussion

This is the first detailed study to measure protein C levels over an extended period of time in patients with severe burns. Assuming that this cohort had normal levels of protein C before injury (which is likely because no patients had an inherited thrombotic condition, infection, pregnancy or condition requiring anticoagulation) the results suggest that a severe burn directly leads to a decrease in plasma protein C. Patients who suffered large burns or deep burns or patients who did not survive admission had the lowest levels of protein C on admission. We used a composite primary outcome called increased support utilisation to identify patients who went on to receive significant emergency, critical care or surgical intervention. Of all the clinical and laboratory data collected, plasma

1667

burns 45 (2019) 1659 –1672

Table 8 – Clinical overview of patients throughout their admission. Standard admission Admitted to, n (%) 13 (22) ICU 46 (78) SBU LOS (days) SBU, mean  SD 16  21 01 ICU, mean  SD Intubated, n (%) Yes 8 (14) 51 (86) No 21 Duration of intubation (days), mean  SD Intravenous fluid administration 2.79  1.81 L Day 0, mean  SD 2.52  1.35 L Day 1, mean  SD 2.31  1.33 L Day 2, mean  SD Surgical management, n (%) 51 (86) Yes 8 (14) No Surgeries, mean  SD 21 Mortality, n (%) 0 (0) Died 59 (100) Survived

Increased support

P value

23 (85) 4 (15)

0.0001

44  40 17  21

<0.0001 <0.0001

15 (55) 12 (45) 11  9

0.0001

5.11  4.16 L 6.86  2.89 L 4.39  1.98 L

0.004 <0.0001 <0.0001

25 (93) 2 (7) 65

0.4949

3 (11) 24 (89)

0.0286

Odds ratio (95% CI)

0.0147

<0.0001

P values formatted to bold font are all values <0.05.

protein C level on admission was most strongly associated with increased support. Twenty-seven patients (31%) had increased support utilisation throughout admission and among these patients, the initial magnitude of loss of functional protein C was proportionate to the risk of subsequently using increased support throughout admission. In order to quantify a protein C cut-off value that was associated with a reduced risk of using increased support, we applied Youden’s index from protein C’s ROC curve. This

Table 9 – Exploratory univariable regression for possible predictors of increased support in patients with burns on admission to hospital. Covariate Age (years) Sex (female) Burn size (TBSA %) Burn depth Partial thickness Full thickness Presence of inhalational injury Protein C (%) CRP (mg/L) Neutrophils 109 Lymphocytes 109 Platelets x109 INR APTT (seconds) Lactate (mmol/L) Creatinine (umol/L) Albumin (g/L)

Odds ratio (95% CI)

P value

1.010 (0.986–1.035) 1.353 (0.488–3.756) 1.100 (1.042–1.160)

0.419 0.561 0.001

Reference 8.330 (2.957–23.466) 8.081 (2.841–22.987) 0.899 (0.849–0.953) 0.986 (0.971–1.003) 1.230 (1.105–1.368) 0.893 (0.580–1.374) 1.003 (0.998–1.008) 2.471 (0.136–44.988) 1.020 (0.916–1.134) 1.964 (1.135–3.399) 1.031 (1.003–1.059) 0.989 (0.922–1.062)

<0.0005

P values formatted to bold font are all values <0.05.

Table 10 – Factors on admission which predict increased support in patients with severe burns (multivariable regression).

<0.0005 <0.0005 0.098 <0.0005 0.606 0.285 0.541 0.723 0.016 0.028 0.769

Burn size (% TBSA) Burn depth Partial thickness Full thickness Protein C (%) Neutrophil count 109

P value

1.252 (1.025–1.530)

0.027

Reference 80.499 (1.569–4129.248) 0.825 (0.698–0.977) 1.532 (0.950–2.473)

0.029 0.025 0.080

showed that normal levels of protein C (70%) optimally improved the sensitivity and specificity of the multivariable regression model. This suggests that a patient with normal protein C levels may be protected from using increased support. Burn size, burn depth, neutrophil count on admission and protein C level on admission contributed together to predict increased support utilisation with high sensitivity, specificity and accuracy. Including protein C level in the multivariable regression predicted increased support utilisation with an additional 20.3% sensitivity, 6.1% specificity, 10.6% accuracy, and almost doubled Nagelkerke’s R2 value explaining 33.8% more variance, compared to when protein C levels were not considered. Other studies have shown that inhalational injury influences patient outcomes [77–79] however when we included it in the stepwise analysis, it lost significance and did not contribute to the multivariable regression model. This may be due to the relatively low sample size or the strength of the model. The findings of this study may be explained by the excessive inflammation and coagulation that occurs after a burn. Based on other studies of protein C in patients and animals with an excessive systemic inflammatory response, and an understanding of burn pathophysiology, we propose three main factors which could contribute to low protein C in burns: impaired synthesis, increased degradation and excessive consumption. The first proposed mechanism is impaired synthesis. Protein C is synthesised largely by the liver, and a burn acutely blunts synthesis of the hepatically produced proteins a-1 acid glycoprotein, C-reactive protein and haptoglobin [80]. The synthesis of protein C may also be suppressed in the same manner after a burn. Protein C is also synthesised by epidermal keratinocytes and vascular endothelium [24], so large burns with significant epidermal and endothelial damage causes direct traumatic loss of functioning protein C-producing cells. Furthermore, pro-inflammatory cytokines such as TNF-a and IL-1 effectively shut off the protein C system [81]. In non-survivors after severe burns, at 24 h after presentation to hospital, IL-1 and TNF-a concentrations are >1000% higher than normal levels [29]. TNF-a suppresses protein C activation [82] while IL-1 suppresses protein C expression [83] and because both cytokines are abundant immediately after a burn, the protein C system is shut off. The second proposed mechanism is increased degradation. Our results show that the neutrophil count is elevated immediately after a burn (mean neutrophil count on admission 10.7  6.1  10 [9], normal range 2–8  10 [9]), and protein C is very sensitive to complete degradation by brief exposures to low concentrations of the proteolytic enzyme neutrophil

1668

burns 45 (2019) 1659 –1672

elastase [84]. This is secreted by neutrophils in response to sepsis and burns [85]. The third proposed mechanism is excessive consumption due to increased thrombin generation following a burn. Thrombin generation is significantly accelerated, and to a greater than normal magnitude, following a burn [86]. Thrombin adherence to the protein C complex is the trigger for the activation of protein C to APC [23]. APC is then rapidly depleted through neutralisation or adherence to plasminogen activator inhibitor 1 [87] and it is plausible that protein C levels decrease when APC is continuously exhausted on a large scale in response to the widespread thrombin generation seen following a burn. The slow return to baseline levels can be explained by the brief half-life of protein C, relative to the slow nature of protein C synthesis. A study of protein C deficient patients who administered a protein C rich substrate of either fresh frozen plasma (FFP) or prothrombin concentrate (PCC), showed that protein C levels declined steadily to baseline deficient levels after approximately 12 and 48 h respectively, with half-lives of 7.8 and 7.4 h respectively [30]. Another study of protein C deficient patients found the half-life of protein C following PCC administration to be 7.9  1.6 h [88]. The effect of reduced protein C system activity is to lose appropriate modulation of inflammation and coagulation. Regarding inflammation, protein C modulates the inflammatory response caused by tissue necrosis and apoptosis [89]. After tissue injury, protein C, which is circulating freely in plasma, moves to bind to the vascular endothelial protein C receptor (EPCR) where it is cleaved by thrombin to form APC [90] and this activates nearby protease-activated receptors (PARs) and the tyrosine kinase-associated receptor, Tie 2 [91,92]. PAR-1, PAR-2 and Tie2 activation [92] leads to a reduction in NF-kB synthesis and nuclear translocation followed by an overall reduction in production of inflammatory mediators [89]. Chemotaxis and cytokine production are then suppressed leading to both decreased leukocyte adhesion and activation and decreased capillary permeability [89]. Multiple studies show the association between excessive circulating pro-inflammatory cytokine levels, pro-inflammatory enzyme levels and poor outcomes after burns [19–22]. Additionally, excessive capillary permeability following a burn [93,94] can lead to catastrophic outcomes such as abdominal compartment syndrome and acute pulmonary oedema [95]. In contrast, a properly functioning protein C system effectively modulates this pro-inflammatory response [23,89,96] and stabilises the endothelium preventing capillary permeability [31,90,97,98]. In unwell patients with normal protein C levels, a drop in protein C levels precedes significant clinical deterioration. In patients who are febrile after chemotherapy-induced neutropaenia, a significant decrease in protein C levels reliably precedes the onset of sepsis [99]. Similarly, in patients who are critically ill from various causes, a significant decrease in protein C levels precedes the onset of severe pancreatitis and multi-organ failure [100]. This suggests that protein C provides a crucial amount of support for immune-mediated processes, and if it is consumed or depleted, unregulated inflammatory processes can take hold. Regarding coagulation, protein C appropriately inhibits factor Va and factor VIIIa, two cofactors which respectively inhibit factor X and prothrombin so that fibrin clots form

appropriately [25]. A detailed study of nine patients with severe burns (four were non-survivors) measured multiple components of the coagulation system including coagulation factors, anticoagulant proteins, the fibrinolytic markers and inflammatory cytokines regularly over the first 96 h of a hospital admission. The only components showing a significant difference between non-survivors and survivors on admission were protein C activity levels (56  18% for nonsurvivors vs. 82  9% for survivors, P = 0.04) and factor VIII activity levels (518  182% for non-survivors vs. 305  148% for survivors, P < 0.05) [29]. Given that protein C inhibits factor VIIIa, this fits with our findings. Given that severe burns lead to an early hypercoagulable state [101] this may be mediated by loss of activity of the protein C system. Given this initial loss of anticoagulant activity, plus the fact that burns are initially treated with high volumes of intravenous crystalloid solutions [43–45] (which further exacerbates coagulopathy) [102–105], the late increases in procoagulant components D-dimer and fibrin [29] and the devastating effects of disseminated intravascular coagulation (DIC) seen after a burn [106], significant attention should be paid to the early reduction of functional protein C levels in burns. The findings of this study give rise to two new concepts about the potential clinical utility of protein C in burns. Firstly, measuring functional protein C levels provides useful prognostic information in patients with burns, such that it may become an adjunct to the clinical examination. Burns clinicians rely heavily on the initial clinical examination to determine early treatment decisions however this method yields unreliable prognostic information [16]. Size estimations [107], depth estimations [108–110] and the use of features on history and examination to diagnose inhalational injury [111] each have shortcomings in reliability and accuracy. Yet they determine treatment that carries an appreciable risk of harm. Further analysis of the role of protein C in burns may allow for the development of superior predictive scores to those currently used. The second concept is that protein C deficiency caused by a burn is harmful and this deficiency could be corrected by replacing it. FFP contains a physiological amount [112] (87  15 units/dL) of protein C [30] and is used by several resuscitation protocols. A well-designed randomised trial compared the use of the Slater formula, which predominantly uses FFP, to the Parkland formula, which predominantly uses Hartmann’s solution. The FFP group achieved the desired urine output with nearly half the fluid of the Hartmann’s group (140 mL/kg/day vs. 260 mL/kg/day, P = 0.005), with significant and clinically important differences favouring FFP with regards to weight gain, peak intra-abdominal pressures, peak inspiratory pressures and base deficit clearance [113]. These findings are consistent with other studies [114,115]. Although there are rare reports of lung injury following FFP transfusion [116], and FFP is more costly than crystalloids or albumin [117], a retrospective review of 5 years’ experience with FFP-based resuscitation deemed FFP to be safe and effective for fluid resuscitation [118]. Future studies of fluid administration in patients with severe burns could help clarify the potential superiority of FFP over other fluid types during acute resuscitation, and whether this potential superiority is due to the influence of protein C or to another property of colloids.

burns 45 (2019) 1659 –1672

5.

Conclusion

In summary, in patients with severe burns, protein C is low on admission to hospital and returns gradually to a steadystate after six days. The initial low measurement can be plausibly explained by impaired synthesis, increased degradation and excessive consumption of protein C following a burn. A plasma protein C level on the day of admission, together with burn size, burn depth, and neutrophil count on the day of admission predicted with high accuracy, sensitivity and specificity the need for increased support throughout admission after a severe burn. It may be a useful biomarker of burn severity.

Funding Funding for this study was supported by the Ramsay Research and Teaching Fund.

Conflict of interest Christopher Jackson, John Vandervord, Gregory Fulcher and Aruna Wijewardena are shareholders in a company undertaking a trial of a 3K3A-APC in diabetic ulcers. For other authors, no competing financial interests exist. The content of this article was expressly written by the authors listed. No ghostwriters were used to write this article.

Acknowledgements We thank Dr. Nancy Huang and Mr. Andrew Zimmerman for their assistance with data collection, and to the staff of the Emergency Department, Intensive Care Department and the Severe Burns Unit at Royal North Shore Hospital for their assistance in this project. We also thank Associate Professor Meilang Xue for technical support and the biostatisticians at The Kolling Institute, in particular Dr. Danny Kim for his advice on regression models. We also thank David Bellingham for his advice on data management and Dr. Boris Waldman for his advice on testing outcome validity. REFERENCES

[1] Girardot T, Rimmelé T, Venet F, Monneret G. Apoptosisinduced lymphopenia in sepsis and other severe injuries. Apoptosis 2017;22:295–305. [2] Lopez ON, Cambiaso-Daniel J, Branski LK, Norbury WB, Herndon DN. Predicting and managing sepsis in burn patients: current perspectives. Ther Clin Risk Manag 2017;13:1107. [3] Heideman M. The effect of thermal injury on hemodynamic, respiratory, and hematologic variables in relation to complement activation. J Trauma 1979;19:239–47. [4] Bartlett RH, Fong SW, Marrujo G, Hardeman J, Anderson W. Coagulation and platelet changes after thermal injury in man. Burns 1981;7:370–7.

1669

[5] Smith CA, Gosselin RC, Utter GH, Galante JM, Young JB, Scherer LA, et al. Does saline resuscitation affect mechanisms of coagulopathy in critically ill trauma patients? An exploratory analysis. Blood Coagul Fibrinolysis 2015;26:250–4. [6] Butt MA, Coulson A, Hull J, Ho T. Early treatment strategies in sepsis. Br Med J 2008;336:521. [7] Strassle PD, Williams FN, Napravnik S, et al. Improved survival of patients with extensive burns: trends in patient characteristics and mortality among burn patients in a tertiary care burn facility, 2004-2013. J Burn Care Res 2017;38:187–93. [8] Asaria P, Elliott P, Douglass M, et al. Acute myocardial infarction hospital admissions and deaths in England: a national follow-back and follow-forward record-linkage study. Lancet Public Health 2017;2:e191–201. [9] Giri S, Pathak R, Aryal MR, Karmacharya P, Bhatt VR, Martin MG. Lack of “Weekend effect” on mortality for pulmonary embolism admissions in 2011: data from nationwide inpatient sample. Int J Cardiol 2015;180:151–3. [10] Duke JM, Boyd JH, Randall SM, Wood FM. Long term mortality in a population-based cohort of adolescents, and young and middle-aged adults with burn injury in Western Australia: a 33-year study. Accid Anal Prev 2015;85:118–24. [11] Duke JM, Boyd JH, Rea S, Randall SM, Wood FM. Long-term mortality among older adults with burn injury: a populationbased study in Australia. Bull World Health Organ 2015;93:400–6. [12] Griffiths C, Guest E, White P, et al. A systematic review of patient-reported outcome measures used in adult burn research. J Burn Care Res 2017;38:e521–45. [13] Friedstat JS, Ryan CM, Gibran N. Outcome metrics after burn injury: from patient-reported outcome measures to valuebased health care. Clin Plast Surg 2017;44:911–5. [14] Osborne CL, Petersson C, Graham JE, et al. The Burn Model Systems outcome measures: a content analysis using the international classification of functioning, disability, and health. Disabil Rehabil 2017;39:2584–93. [15] Saly D, Yang A, Triebwasser C, et al. Approaches to predicting outcomes in patients with acute kidney injury. PLoS One 2017;12:e0169305. [16] Hussain A, Dunn KW. Predicting length of stay in thermal burns: a systematic review of prognostic factors. Burns 2013;39:1331–40. [17] Cabral L, Afreixo V, Almeida L, Paiva JA. The use of procalcitonin (PCT) for diagnosis of Sepsis in burn patients: a meta-analysis. PLoS One 2016;11:e0168475. [18] Cabral L, Afreixo V, Meireles R, et al. Checking procalcitonin suitability for prognosis and antimicrobial therapy monitoring in burn patients. Burns Trauma 2018;6:10. [19] Shelhamer MC, Rowan MP, Cancio LC, et al. Elevations in inflammatory cytokines are associated with poor outcomes in mechanically ventilated burn patients. J Trauma Acute Care Surg 2015;79:431–6. [20] Finnerty CC, Jeschke MG, Qian WJ, et al. Determination of burn patient outcome by large-scale quantitative discovery proteomics. Crit Care Med 2013;41:1421–34. [21] Hastbacka J, Freden F, Hult M, et al. Matrix metalloproteinases -8 and -9 and tissue inhibitor of metalloproteinase-1 in burn patients. A prospective observational study. PLoS One 2015;10:e0125918. [22] Huang L, Yao Y, Dong N, et al. Prognostic significance of plasma gelsolin in severe burn patients with sepsis. Zhonghua Shao Shang Za Zhi 2016;32:77–81. [23] Mosnier LO, Zlokovic BV, Griffin JH. The cytoprotective protein C pathway. Blood 2007;109:3161–72. [24] Xue M, Campbell D, Jackson CJ. Protein C is an autocrine growth factor for human skin keratinocytes. J Biol Chem 2007;282:13610–6.

1670

burns 45 (2019) 1659 –1672

[25] Dahlbäck B, Villoutreix BO. Regulation of blood coagulation by the protein C anticoagulant pathway: novel insights into structure-function relationships and molecular recognition. Arterioscler Thromb Vasc Biol 2005;25:1311–20. [26] Miletich J, Sherman L, Broze G. Absence of thrombosis in subjects with heterozygous protein C deficiency. N Engl J Med 1987;317:991–6. [27] Fernandez JA, Petaja J, Gruber A, Griffin JH. Activated protein C correlates inversely with thrombin levels in resting healthy individuals. Am J Hematol 1997;56:29–31. [28] Marlar RA, Gausman JN. Laboratory testing issues for protein C, protein S, and antithrombin. Int J Lab Hematol 2014;36:289–95. [29] Tejiram S, Brummel-Ziedins KE, Orfeo T, et al. In-depth analysis of clotting dynamics in burn patients. J Surg Res 2016;202:341–51. [30] Marlar RA, Sills RH, Groncy PK, Montgomery RR, Madden RM. Protein C survival during replacement therapy in homozygous protein C deficiency. Am J Hematol 1992;41:24– 31. [31] Whitmont K, Fulcher G, Reid I, et al. Low circulating protein C levels are associated with lower leg ulcers in patients with diabetes. BioMed Res Int 20132013:. [32] Esechie A, Enkhbaatar P, Nakano YY, Traber L, Herndon D, Traber DL. Comparison of plasma concentrations of antithrombin (AT) and protein c (PC) in burn and smoke inhalation injury and sepsis. FASEB J 2006;20:A286–7. [33] Lo S, Lai W, Kwok F. Protein C and Protein S levels in some burns patients. Burns 1994;20:186–7. [34] Von Elm E, Altman DG, Egger M, et al. The strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement: guidelines for reporting observational studies. PLoS Med 2007;4:e296. [35] Innovation AfC. NSW burn transfer guidelines. In: NGDo Health, editor. Statewide burn injury service. . [36] Demling RH, Wong C, Jin L, Hechtman H, Lalonde C, West K. Early lung dysfunction after major burns: role of edema and vasoactive mediators. J Trauma 1985;25:959–66. [37] Ennis JL, Chung KK, Renz EM, et al. Joint Theater Trauma System implementation of burn resuscitation guidelines improves outcomes in severely burned military casualties. J Trauma 2008;64:S146–51 Discussion S51–52. [38] Malbrain ML, Cheatham ML, Kirkpatrick A, et al. Results from the international conference of experts on intra-abdominal hypertension and abdominal compartment syndrome. I. Definitions. Intensive Care Med 2006;32:1722–32. [39] Ivy ME, Atweh NA, Palmer J, Possenti PP, Pineau M, D’aiuto M. Intra-abdominal hypertension and abdominal compartment syndrome in burn patients. J Trauma Acute Care Surg 2000;49:387–91. [40] Parra MW, Al-Khayat H, Smith HG, Cheatham ML. Paracentesis for resuscitation-induced abdominal compartment syndrome: an alternative to decompressive laparotomy in the burn patient. J Trauma Acute Care Surg 2006;60:1119–21. [41] Dulhunty JM, Boots RJ, Rudd MJ, Muller MJ, Lipman J. Increased fluid resuscitation can lead to adverse outcomes in major-burn injured patients, but low mortality is achievable. Burns 2008;34:1090–7. [42] Lund T, Onarheim H, Reed RK. Pathogenesis of edema formation in burn injuries. World J Surg 1992;16:2–9. [43] Baxter CR, Shires T. Physiological response to crystalloid resuscitation of severe burns. Ann N Y Acad Sci 1968;150:874–94. [44] Baker RHJ, Akhavani MA, Jallali N. Resuscitation of thermal injuries in the United Kingdom and Ireland. J Plast Reconstr Aesthetic Surg 2007;60:682–5. [45] Gordon M, Marvin J, Greenfield E, et al. Regional and institutional variation in burn care. J Burn Care Rehabil 1995;16:85–90.

[46] Engrav LH, Colescott PL, Kemalyan N, et al. A biopsy of the use of the Baxter formula to resuscitate burns or do we do it like Charlie did it? J Burn Care Rehabil 2000;21:91–5. [47] Cartotto RC, Innes M, Musgrave MA, Gomez M, Cooper AB. How well does the Parkland formula estimate actual fluid resuscitation volumes? J Burn Care Rehabil 2002;23:258–65. [48] Friedrich JB, Sullivan SR, Engrav LH, et al. Is supra-Baxter resuscitation in burn patients a new phenomenon? Burns 2004;30:464–6. [49] Klein MB, Hayden D, Elson C, et al. The association between fluid administration and outcome following major burn: a multicenter study. Ann Surg 2007;245:622–8. [50] Ivy ME, Atweh NA, Palmer J, Possenti PP, Pineau M, D’Aiuto M. Intra-abdominal hypertension and abdominal compartment syndrome in burn patients. J Trauma Inj Infect Crit Care 2000;49:387–91. [51] Oda J, Yamashita K, Inoue T, et al. Resuscitation fluid volume and abdominal compartment syndrome in patients with major burns. Burns 2006;32:151–4. [52] Brandstrup B, Tønnesen H, Beier-Holgersen R, et al. Effects of intravenous fluid restriction on postoperative complications: comparison of two perioperative fluid regimens: a randomized assessor-blinded multicenter trial. Ann Surg 2003;238:641. [53] Liu NT, Fenrich CA, Serio-Melvin ML, Peterson WC, Cancio LC, Salinas J. The impact of patient weight on burn resuscitation. J Trauma Acute Care Surg 2017;83:S112–9. [54] Maan ZN, Frew Q, Din AH, et al. Burns ITU admissions: length of stay in specific levels of care for adult and paediatric patients. Burns 2014;40:1458–62. [55] Peck MD, Mantelle L, Ward CG. Comparison of length of hospital stay to mortality rate in a regional burn center. J Burn Care Rehabil 1996;17:39–44. [56] Lemeshow S, Klar J, Teres D, et al. Mortality probability models for patients in the intensive care unit for 48 or 72 hours: a prospective, multicenter study. Crit Care Med 1994;22:1351–8. [57] Garside TL, Lee RP, Delaney A, Milliss D. Clinical practice variation in acute severe burn injury. Anaesth Intensive Care 2018;46:321–5. [58] Hodgman EI, Subramanian M, Wolf SE, et al. The effect of illicit drug use on outcomes following burn injury. J Burn Care Res 2017;38:e89–94. [59] Blay B, Thomas S, Coffey R, Jones L, Murphy CV. Low vitamin d level on admission for burn injury is associated with increased length of stay. J Burn Care Res 2017;38:e8–e13. [60] Issler-Fisher AC, Fakin RM, Fisher OM, et al. Microbiological findings in burn patients treated in a general versus a designated intensive care unit: effect on length of stay. Burns 2016;42:1805–18. [61] van Langeveld I, Gagnon RC, Conrad PF, et al. Multiple-drug resistance in burn patients: a retrospective study on the impact of antibiotic resistance on survival and length of stay. J Burn Care Res 2017;38:99–105. [62] Pandharipande PP, Girard TD, Jackson JC, et al. Long-term cognitive impairment after critical illness. N Engl J Med 2013;369:1306–16. [63] Moitra VK, Guerra C, Linde-Zwirble WT, Wunsch H. Relationship between ICU length of stay and long-term mortality for elderly ICU survivors. Crit Care Med 2016;44:655. [64] Minne L, Abu-Hanna A, de Jonge E. Evaluation of SOFA-based models for predicting mortality in the ICU: a systematic review. Crit Care 2008;12:R161. [65] Israel JS, Greenhalgh DG, Gibson AL. Variations in burn excision and grafting: a survey of the american burn association. J Burn Care Res 2017;38:e125–32. [66] Nosanov LB, McLawhorn MM, Hassan L, et al. Graft loss: review of a single burn center’s experience and proposal of a graft loss grading scale. J Surg Res 2017;216:185–90.

burns 45 (2019) 1659 –1672

[67] Abbott TEF, Ahmad T, Phull MK, et al. The surgical safety checklist and patient outcomes after surgery: a prospective observational cohort study, systematic review and metaanalysis. Br J Anaesth 2017;120:146–55. [68] Urbach DR, Govindarajan A, Saskin R, Wilton AS, Baxter NN. Introduction of surgical safety checklists in Ontario, Canada. N Engl J Med 2014;370:1029–38. [69] Li H, Yao Z, Tan J, et al. Epidemiology and outcome analysis of 6325 burn patients: a five-year retrospective study in a major burn center in Southwest China. Sci Rep 2017;7:46066. [70] Gangemi EN, Gregori D, Berchialla P, et al. Epidemiology and risk factors for pathologic scarring after burn wounds. Arch Facial Plast Surg 2008;10:93–102. [71] Mason SE, Noel-Storr A, Ritchie CW. The impact of general and regional anesthesia on the incidence of post-operative cognitive dysfunction and post-operative delirium: a systematic review with meta-analysis. J Alzheimers Dis 2010;22:S67–79. [72] Desborough J. The stress response to trauma and surgery. Br J Anaesth 2000;85:109–17. [73] Guo Z, Liu J, Li J, et al. Postoperative delirium in severely burned patients undergoing early escharotomy: incidence, risk factors, and outcomes. J Burn Care Res 2017;38:e370–6. [74] Royal North Shore Hospital LabInfo Test Directory. 2017 [Accessed 2 December 2017, at http://www.palms.com.au/ php/labinfo/info_index.php?tab=4]. [75] Rizzo JA, Rowan MP, Driscoll IR, Chan RK, Chung KK. Perioperative temperature management during burn care. J Burn Care Res 2017;38:e277–83. [76] Hosmer [455_TD$DIFF]Jr D, Lemeshow S, Sturdivant R. Applied logistic regression. 3rd ed. New York: John Wiley & Sons. Inc; 2013. [77] Shirani KZ, Pruitt [456_TD$DIFF]Jr BA, Mason Jr AD. The influence of inhalation injury and pneumonia on burn mortality. Ann Surg 1987;205:82. [78] Smith DL, Cairns BA, Ramadan F, et al. Effect of inhalation injury, burn size, and age on mortality: a study of 1447 consecutive burn patients. J Trauma 1994;37:655–9. [79] Blot S. Development and validation of a model for prediction of mortality in patients with acute burn injury: the Belgian Outcome in Burn Injury Study Group. Br J Surg 2008;96:111–7. [80] Thomas S, Wolf SE, Chinkes DL, Herndon DN. Recovery from the hepatic acute phase response in the severely burned and the effects of long-term growth hormone treatment. Burns 2004;30:675–9. [81] Levi M, van der Poll T. Inflammation and coagulation. Crit Care Med 2010;38:S26–34. [82] Nawroth PP, Stern DM. Modulation of endothelial cell hemostatic properties by tumor necrosis factor. J Exp Med 1986;163:740–5. [83] Yamamoto K, Shimokawa T, Kojima T, Loskutoff DJ, Saito H. Regulation of murine protein C gene expression in vivo: effects of tumor necrosis factor-alpha, interleukin-1, and transforming growth factor-beta. Thromb Haemost 1999;82:1297–301. [84] Eckle I, Seitz R, Egbring R, Kolb G, Havemann K. Protein C degradation in vitro by neutrophil elastase. Biol Chem Hoppe Seyler 1991;372:1007–13. [85] Kaufman T, Magosevich D, Moreno MC, et al. Nucleosomes and neutrophil extracellular traps in septic and burn patients. Clin Immunol 2017;183:254–62. [86] Wiegele M, Schaden E, Koch S, Bauer D, Krall C, Adelmann D. Thrombin generation in patients with severe thermal injury. Burns 2019;45:54–62. [87] Hazelzet JA. Levels of protein C and activated protein C: what do they mean? Crit Care 2006;10:126. [88] Riess H, Binsack T, Hiller E. Protein C antigen in prothrombin complex concentrates: content, recovery and half life. Blut 1985;50:303–6.

1671

[89] Esmon CT. Protein C anticoagulant system—antiinflammatory effects. Semin Immunopathol 2012;34:127–32. [90] Jackson CJ, Xue M. Activated protein C: an anticoagulant that does more than stop clots. Int J Biochem Cell Biol 2008;40:2692–7. [91] Julovi SM, Xue M, Dervish S, Sambrook PN, March L, Jackson CJ. Protease activated receptor-2 mediates activated protein C-induced cutaneous wound healing via inhibition of p38. Am J Pathol 2011;179:2233–42. [92] Xue M, Chow SO, Dervish S, Chan YKA, Julovi SM, Jackson CJ. Activated protein C enhances human keratinocyte barrier integrity via sequential activation of epidermal growth factor receptor and tie2. J Biol Chem 2011;286:6742–50. [93] Pitt RM, Parker JC, Jurkovich GJ, Taylor AE, Curreri PW. Analysis of altered capillary pressure and permeability after thermal injury. J Surg Res 1987;42:693–702. [94] Harms BA, Bodai BI, Kramer GC, Demling RH. Microvascular fluid and protein flux in pulmonary and systemic circulations after thermal injury. Microvasc Res 1982;23:77–86. [95] Demling RH. The burn edema process: current concepts. J Burn Care Rehabil 2005;26:207–27. [96] Pina-Canseco MS, Páez-Arenas A, Massó F, et al. Protein C activation peptide inhibits the expression of ICAM-1, VAM-1, and interleukin-8 induced by TNF-a in human dermal microvascular endothelial cells. Folia Histochem Cytobiol 2012;50:407–13. [97] Whitmont K, Reid I, Tritton S, et al. Treatment of chronic leg ulcers with topical activated protein C. Arch Dermatol 2008;144:1479–83. [98] Jackson C, Whitmont K, Tritton S, March L, Sambrook P, Xue M. New therapeutic applications for the anticoagulant, activated protein C. Expert Opin Biol Ther 2008;8:1109–22. [99] Mesters RM, Helterbrand J, Utterback BG, et al. Prognostic value of protein C concentrations in neutropenic patients at high risk of severe septic complications. Crit Care Med 2000;28:2209–16. [100] Lindstrom O, Kylanpaa L, Mentula P, et al. Upregulated but insufficient generation of activated protein C is associated with development of multiorgan failure in severe acute pancreatitis. Crit Care 2006;10:R16. [101] Pruitt Jr. BA, Wolf SE. An historical perspective on advances in burn care over the past 100 years. Clin Plast Surg 2009;36:527–45. [102] Ruttmann TG, James MFM, Finlayson J. Effects on coagulation of intravenous crystalloid or colloid in patients undergoing peripheral vascular surgery. Br J Anaesth 2002;89:226–30. [103] Ruttmann TG, James MFM, Lombard EH. Haemodilutioninduced enhancement of coagulation is attenuated in vitro by restoring antithrombin III to pre-dilution concentrations. Anaesth Intensive Care 2001;29:489–93. [104] Ng FJ, Lam CCK, Chan LC. In vivo effect of haemodilution with saline on coagulation: a randomized controlled trial. Br J Anaesth 2002;88:475–80. [105] Boldt J, Haisch G, Suttner S, Kumle B, Schellhase F. Are lactated Ringer’s solution and normal saline solution equal with regard to coagulation? Anesth Analg 2002;94:378–84. [106] Boral BM, Williams DJ, Boral LI. Disseminated intravascular coagulation. Am J Clin Pathol 2016;146:670–80. [107] Tocco-Tussardi I, Presman B, Huss F. Want correct percentage of TBSA burned? Let a layman do the assessment. J Burn Care Res 2018;39:295–301. [108] Monstrey S, Hoeksema H, Verbelen J, Pirayesh A, Blondeel P. Assessment of burn depth and burn wound healing potential. Burns 2008;34:761–9. [109] Watts AM, Tyler MP, Perry ME, Roberts AH, McGrouther DA. Burn depth and its histological measurement. Burns 2001;27:154–60. [110] Devgan L, Bhat S, Aylward S, Spence RJ. Modalities for the assessment of burn wound depth. J Burns Wounds 2006;5:e2.

1672

burns 45 (2019) 1659 –1672

[111] Ching JA, Shah JL, Doran CJ, Chen H, Payne WG, Smith Jr. DJ. The evaluation of physical exam findings in patients assessed for suspected burn inhalation injury. J Burn Care Res 2015;36:197–202. [112] Tait RC, Walker ID, Islam SI, et al. Protein C activity in healthy volunteers—influence of age, sex, smoking and oral contraceptives. Thromb Haemost 1993;70:281–5. [113] O’Mara MS, Slater H, Goldfarb IW, Caushaj PF. A prospective, randomized evaluation of intra-abdominal pressures with crystalloid and colloid resuscitation in burn patients. J Trauma 2005;58:1011–8. [114] Du G-B, Slater H, Goldfarb IW. Influences of different resuscitation regimens on acute early weight gain in extensively burned patients. Burns 1991;17:147–50.

[115] Hoelscher V, Harvin J, Cotton B, Wade C, Huzar T. 218 impact of fresh frozen plasma infusions during resuscitation in thermally-injured patients. J Burn Care Res 201839: S76-S. [116] Jones LM, Deluga N, Bhatti P, Scrape SR, Bailey JK, Coffey RA. TRALI following fresh frozen plasma resuscitation from burn shock. Burns 2017;43:397–402. [117] Cartotto R, Greenhalgh D. Colloids in acute burn resuscitation. Crit Care Clin 2016;32:507–23. [118] Jones L, Brown N, Philips G, et al. Burn resuscitation with fresh frozen plasma: 5 years of experience with the west Penn formula. Austin J Emerg Crit Care Med 2015;2:1018.