SURGICAL INFECTIONS

Volume XX, Number XX, 2020 Original Article

ª Mary Ann Liebert, Inc.
DOI: 10.1089/sur.2020.348

Detection of Infection and Sepsis in Burns

Mark Jason M. Torres,1 Joshua M. Peterson,2, and Steven E. Wolf 2,3

Abstract

Background: Infection is the most frequent complication after severe burns and has a propensity to progress
into sepsis then septic shock and multiple organ dysfunction syndrome (MODS). Improving outcomes in acute
Downloaded by Universiteit Utrecht from www.liebertpub.com at 10/08/20. For personal use only.

burn care depends on early detection of infection to allow prompt interventions. Diagnosis of sepsis in severe
burns is uniquely challenging because otherwise-typical clinical signs are masked by the hypermetabolic state
and systemic inflammation induced by the burn itself. For this reason, burns have historically been excluded
from high-impact studies on the diagnosis and treatment of sepsis.
Methods: This article provides a comprehensive three-fold review of current findings and guidelines pertinent
to the early detection of infection and sepsis in severe burns.
Results: First, evidence-based detection of the most common infections encountered in the burn intensive care
unit is reviewed. Second, we analyze the evolution of the diagnostic criteria for sepsis and the evidence
regarding their utility in severe burns. Last, we examine the development of biomarkers, from procalcitonin to
molecular genomics, for the detection of sepsis.
Conclusions: Although gold standard methods of early detection of sepsis in burn patients have yet to be
identified, improved understanding and appropriate application of the available diagnostic criteria and assays
are paramount to providing effective care of patients with severe burns.

Keywords: burn; burn injury; burn wound; infection; pneumonia; procalcitonin; sepsis; sepsis biomarkers;
septic shock; wound infection

I nfection leading to sepsis, septic shock, and multiple

organ dysfunction syndrome (MODS) is well known as a
leading cause of death after severe burns, accounting for
article, we review the current modalities for the detection of
infections and sepsis in severe burns as well as explore why a
consensus regarding appropriate diagnostic methods has not
50%–60% of in-patient deaths [1,2]. With the growth of yet been achieved.
specialized burn units, implementation of fluid resuscitation,
and early excision and grafting of burn wounds, the historical Detection of Infection
trend of early shock-related death after severe burns has
Among patients admitted to the burn intensive care unit
transitioned to a trend of death from infection and sepsis after
(ICU), infection is the most frequent complication. Pneu-
prolonged hospitalization [3,4]. Burned patients are particu-
monia and wound infections are most common and often are
larly susceptible to wound infections and pneumonia (espe-
nosocomial in origin [5]. For the severely burned, unique
cially after inhalation injury), which can progress ultimately
considerations must be made during evaluation for possible
to sepsis. Establishing a timely diagnosis of sepsis has proved
pneumonia or wound infections.
uniquely challenging in the background of the intense hy-
permetabolic and inflammatory state associated with massive
Pneumonia
injury, confounding traditional clinical signs of sepsis and
greatly hindering the utility of contemporary diagnostic Respiratory infections are the most frequently reported
methods. Moreover, frequent fluctuations in the definition of infections in burned patients, who have the highest overall
sepsis as well as exclusion of burn patients from landmark rate of ventilator-associated pneumonia (VAP) (7.4%) of all
studies on sepsis have muddled investigative efforts. In this ICU sub-populations. Those with inhalation injury have four

1
School of Medicine, 2Department of Surgery, University of Texas Medical Branch, Galveston, Texas, USA.
3
Shriners Hospitals for Children, Galveston, Texas, USA.

1
 2 TORRES ET AL.

times the risk of developing VAP, a state caused by several is emphasized that treatment for pneumonia in patients
factors including prolonged mechanical ventilation with en- meeting the clinical criteria for diagnosis should not be de-
dotracheal tubes, systemic immunosuppression, common layed while awaiting microbiologic confirmation of infec-
development of pulmonary edema, and direct pulmonary tion. In the interim, initial antibiotic therapy may be guided
tissue injury [6]. Specifically, direct tissue injury to the lower by Gram stain results as well as wound culture findings, as the
respiratory tract predisposes to pneumonia through creation causative organism of pneumonia is found to be identical to
of a pro-infectious microenvironment with massive cytokine organisms isolated from wound cultures in approximately
secretion and inflammation and subsequent protein-rich fluid 50% of cases [8].
secretions, bronchial edema, failure of surfactant production
by injured type II pneumocytes, failure of mucous clearance
Wound infection
by injured ciliated cells, and formation of bronchial casts
from cellular sloughing. These pathophysiologic changes Immediately after a severe burn the affected skin surface
precipitate alveolar collapse, atelectasis, and bacterial growth becomes sterilized and remains free of flora for approxima-
leading to VAP, which occurs at a rate of 10%–60% after tely 2 days until flora from adjacent unaffected skin, the naso-
inhalation injury and is the leading cause of death among oropharyngeal tract, lower gastrointestinal tract, and the
burn patients with nosocomial infections, who have a VAP environment proliferate over the wound [11]. Burn eschars
mortality rate of 40% [7]. Risk factors for contracting allow colonization of microorganisms through abundant co-
pneumonia include grade 3 (severe) or grade 4 (massive) agulated protein, elemental co-factors, and protein-rich ex-
inhalation injury on fiberoptic bronchoscopy (FOB), in- udates. The absence of leukocytes in eschar and loss of skin
Downloaded by Universiteit Utrecht from www.liebertpub.com at 10/08/20. For personal use only.

creasing burn size (total body surface area [TBSA] burned barrier function sets the stage for proliferation of virulent
>20%), PaO2/FiO2 ratio of <300 mm Hg on admission, car- pathogens such as Staphylococcus aureus (both methicillin-
boxyhemoglobin concentrations of >10% on admission, resistant and methicillin-sensitive), Pseudomonas aeruginosa,
and a smoking history [8]. Patients are at highest risk of Klebsiella spp., and Acinetobacter baumannii that outcom-
developing VAP within the first 5 days of mechanical ven- pete other organisms and develop as soon as 5–7 days after
tilation (estimated at 3% per day) with a steadily declining injury [12].
risk to 2% and 1% per day on days 5-10 and after day 10, Between the 1950s and early 1980s, initial therapy for burn
respectively [6]. wounds included repetitive submersion hydrotherapy aimed
Traditional clinical signs of pneumonia including pulmo- at breaking down eschars and promoting underlying wound
nary infiltrates on radiographic imaging, fever, leukocytosis, bed granulation prior to autografting while simultaneously
and purulent tracheobronchial secretions have notoriously suppressing infection through application of topical anti-
poor specificity in the diagnosis of VAP. Studies correlating septic agents such as silver sulfadiazine, mafenide acetate,
clinical diagnosis of VAP with post-mortem histologic ex- and silver nitrate. In addition to clinical signs of infection
amination and cultures demonstrate a sensitivity and speci- such as rapid separation of eschar, foci of eschar discolor-
ficity of clinical criteria as only 69% and 75%, respectively ation, suppurative secretions, and hematogenous spread of
[9]. Outside of autopsy, no gold standard exists for the di- organisms (i.e., ecthyma gangrenosum), the gold standard for
agnosis of VAP, complicating both treatment and meaningful wound infection diagnosis during this era was tissue biopsy
research. with histopathologic analysis and quantitative culture; bac-
To address this deficiency, the American Burn Association terial loads of >105 organisms/g signaled invasive infection
(ABA) released guidelines for the diagnosis and treatment of best treated with wound excision (Table 1). In 1981, a land-
VAP in 2009, supporting a microbiologic approach to diag- mark paper by Woolfrey et al. [13] demonstrated marked
nosis via FOB with sampling of pulmonary flora by bronch- discordance of quantitative cultures (44% with a log10 dif-
ioalveolar lavage (BAL), mini-BAL, or protected specimen ference of 2 or greater) when specimens were divided into two
brush sampling as the recommended diagnostic pathway in separate cultures, casting doubt on their utility. Subsequently,
patients who have VAP suspected on the basis of clinical early excision and grafting of burn eschars was shown to
findings [10]. Then, in 2018, the International Society for improve outcomes dramatically in part by preventing wound
Burn Injuries (ISBI) released practice guidelines (part 2) infections through removal of the nutrient-rich eschar and
wherein the clinical and microbiologic diagnoses of pneu- restoration of skin barrier function. Early excision and graft-
monia were defined as noted in Table 1. They affirmed that ing have since become widely implemented as standard of
BAL and subglottic sampling demonstrated equivalency to care throughout the world.
protected specimen brush sampling in intubated burn patients Of late, wound cultures serve for surveillance of wounds
and are the preferred sampling methods [8]. and to acquire qualitative and semi-quantitative data when
Bronchoscopic microbiologic diagnosis of pneumonia in infection is suspected. In the ISBI 2018 guidelines, wound
burn patients has a sensitivity of 73% – 18% with a specificity surfaces should be cleansed with alcohol-based solutions
of 82% – 19%, and direct sampling is postulated also to have prior to sampling, and sampling should be guided by clini-
therapeutic benefits through mechanical relief of bronchial cal signs such as formation of exudates, widespread graft
obstruction by suction of exudates [6]. In fact, patients with loss, or systemic signs of infection. Importantly, surveillance
inhalation injury who develop pneumonia and undergo FOB swabs should not guide initiation of antimicrobial therapy;
have a shorter duration of mechanical ventilation (21 versus however, they provide useful information about organism
28 days), decreased length of ICU stay (35 versus 39 days) susceptibilities that guide decisions regarding antibiotic
and hospital stay (45 versus 49 days), and a lower total cost of coverage. Tissue biopsy with quantitative culture and his-
care ($370,572 versus $473,654). However, no statistically topathologic analysis retain utility in assessing fungal in-
significant mortality benefit has yet been demonstrated [7]. It fections and should be strongly considered when such
 DETECTING BURN INFECTION AND SEPSIS 3

Table 1. International Society for Burn Injuries (ISBI) 2018 Diagnostic Criteria
for Pneumonia and Wound Infection
Pneumonia
Clinical diagnosis
Two of the following are present:
1. Chest radiograph with a new and persistent infiltrate, consolidation, or cavitation
2. Sepsis
3. Recent change in sputum or purulence in the sputum

Microbiologic modifier: Clinical diagnosis Pathogen isolateda

Confirmed + +
Probable + -
Possibleb - +

Wound infection

Extent Bacterial load Local host response? Systemic host response? Wound healing slowed?
Downloaded by Universiteit Utrecht from www.liebertpub.com at 10/08/20. For personal use only.

Colonization W: <105 CFU/g - - -

Local infection W: >105 CFU/g + - +
Invasive infection W, T: >105 CFU/g + + +
Cellulitis Clinical diagnosisc + +/- +
a
Tracheal aspirate with ‡105 organisms, or bronchoalveolar lavage with ‡104 organisms, or protected specimen brush with ‡103
organisms.
b
Abnormal chest radiograph with uncertain cause and low or moderate clinical suspicion.
c
Adjacent unburned skin is warm, indurated, erythematous, tender.
+ = present, - = absent, CFU = colony-forming units; T = deep tissues; W = wound.

infection is suspected, as invasive fungal infections can stated aim of improving bedside diagnosis rather than
progress rapidly with high rates of death [8]. facilitating precise parameters amenable to investigative
efforts [18,19].
Detection of Sepsis Although MODS and the Sequential Organ Failure
Assessment (SOFA) score were referenced directly by
The progression of infection to sepsis, septic shock, and
Sepsis-2, the focus of the definition rested on SIRS and
MODS remains the leading cause of death in both adult
identification of infection. Importantly, the clinical studies
(50%–84%) and pediatric (55%) burn patients [14]. In 2016,
of sepsis on which these definitions and criteria were con-
the Sepsis-3 criteria defined sepsis as ‘‘a life-threatening
structed excluded burned patients [1,8].
organ dysfunction caused by a dysregulated host response to
In 2007, the ABA convened 23 experts to devise consensus
infection’’ [15]. Unlike many other causes of sepsis, its clinical
definitions specific to burn sepsis, arguing that the Sepsis-2
diagnosis in severe burns is nuanced and elusive, clouded by a
criteria were too inclusive to be applicable to severe burns and
storm of inflammatory cytokines known as damage-associated
did not account for the hypermetabolic state. It made several
molecular patterns (DAMPs) and the pathophysiologic hy-
changes as defined in Table 2, notably the presence of a large
permetabolic response they create. This confounding hyper-
burn (>20% TBSA), establishment of higher thresholds of SIRS
metabolic response mimics systemic inflammatory response
parameters, abandonment of severe sepsis, and addition of
syndrome (SIRS) patterns with tachycardia, tachypnea, ele-
thrombocytopenia, hyperglycemia, and intolerance of enteral
vated body temperatures, and leukocytosis, which develop
feeding as triggers for concern [20]. Furthermore, emphasis was
reliably as a pathophysiologic response to large (>40%)
placed on burn sepsis being defined as a change in patient status.
TBSA burns and persist for more than a year after injury [16].
A subsequent retrospective study by Mann-Salinas et al.
The development of diagnostic criteria and assays for the
evaluated the effectiveness of the ABA criteria in adult burn
accurate and timely diagnosis of sepsis has been a subject of
patients and found that they were predictive of sepsis one day
tremendous investigative effort for more than three decades.
prior to the drawing of blood for culture with an area under the
receiver operator curve (AUROC) of 0.619. Using a statistical
Evolution in sepsis definitions and diagnostic criteria
regression model, they were able to define six sepsis predic-
Prior to 1991, no formal or rigid diagnostic criteria for sepsis tors that improved the AUROC to 0.775 but which did not
existed. To improve the quality of epidemiologic and investi- prove to be accurate in prospective studies [21,22].
gative research, the Society of Critical Care Medicine and the Again in 2016 as a product of the Surviving Sepsis Cam-
American College of Chest Physicians developed a definition paign, Sepsis-3 guidelines revised the definitions, abolishing
now known as Sepsis-1 (Table 2) [17]. Ten years later, Sepsis- SIRS and severe sepsis and incorporating host organ response
2 expanded SIRS to include a vast array of possible parameters into the definition. The SOFA was formally incorporated, with
with infection only documented or suspected, with the openly a score of 2 or greater requisite for diagnosis [15]. Despite
 Downloaded by Universiteit Utrecht from www.liebertpub.com at 10/08/20. For personal use only.

Table 2. Evolution of Adult Sepsis Diagnostic Criteria and Definitions

Criteria Sepsis-3 (2016) ABA burn sepsis (2007) Sepsis-2 (2001) Sepsis-1 (1991)
Sepsis 1. Infectiona 1. Burn injury >20% TBSA 1. Infectiona 1. Infectiona
2. Acute change in SOFA 2. Documented infection 2. Someb of the following: 2. SIRS
j
score ‡2 Positive cultures, or General
j
Tissue source, or T > 38.3C or <36C; HR >90 min-1; VR
j
Clinical response to >20 min-1; AMS; significant edema or
SOFA score with 0–4-point antibiotics fluid retention; BG >120 mg_dL-1 (no SIRS defined as ‡2 of the
range: 3. ‘‘Triggers’’ ‡3 of these: diabetes) following:
j j
Respiration PaO2/FiO2: ‡400 T > 39C or <36.5C Inflammatory T > 38C or <36C
j j
to <100 HR >110 min-1 WBC >12,000 mm-3 or <4000 mm-3 or HR >90 min-1
j j
Coagulation PLT: >150,000 RR >22 min-1 or PaCO2 >10% immature forms; CRP >2 SD VR >20 min-1
to <20,000 mm-3 < 32 mm Hg above normal; PCT >2 SD above or PaCO2 < 32 mm Hg
j j
Liver Total bilirubin: <1.2 to PLT <100,000 mm-3 normal WBC >12,000 or
>12.0 mg_dL-1 (‡3 d after resuscitation) Hemodynamic <4,000 mm-3 or >10%
j
Cardiovascular MAP: ‡70 to BG >200 mg_dL-1; SBP <90 mm Hg, MAP <70, or SBP immature forms
‡70 mm Hg with >7 units_h-1 intravenous decrease >40 mm Hg; SVO2 > 70%;
vasopressors insulin, or >25% dose CI >3.5 L_min-1_m-2
CNS GCS: 15 to <6 increase in 24 h. Organ dysfunction
j
Renal Cr <1.2 to Intolerance of enteral PaO2/FiO2 < 300; UOP
>5.0 mg_dL-1; UOP >500 feeding for >24 h <0.5 mL_kg-1_hr-1; Cr elevation
to <200 mL_d-1 (abdominal distention, >0.5 mg_dL-1; INR >1.5 or aPTT

4
diarrhea >2.5 L_d-1) >60 sec; bowel sounds absent; PLT
<100,000 mm-3; total bilirubin
>4 mg_dL-1
Tissue perfusion
Lactate >1 mmol_L-1; decreased capillary
refill; mottling

Severe sepsis [abandoned] [abandoned] 1. Sepsis 1. Sepsis

2. Organ dysfunction 2. Hypotension,c
(SOFA or Marshall MOD score hypoperfusion, or organ
recommended) dysfunction

Septic shock 1. Sepsis 1. Sepsis 1. Sepsis 1. Sepsis

2. Vasopressors given for 2. Lactate >4 mmol_L-1 2. Hypotensionc 2. Hypotensionc
MAP <65 mm Hg 3. Organ dysfunction 3. Hypoperfusion or organ
3. Lactate >2 mmol_L-1 (SOFA or Marshall MOD score dysfunction
recommended)
aPTT = activated partial thromboplastin time; AMS = altered mental status; BG = blood glucose; CI = cardiac index; CNS = central nervous system; Cr = creatinine; CRP = C-reactive protein;
FiO2 = fraction of inspired oxygen; GCS = Glasgow Coma Scale; HR = heart rate; INR = International Normalized Ratio; MAP = mean arterial pressure; PCT = procalcitonin; PaO2 = arterial partial
pressure of oxygen; PLT = platelets; SBP = systolic blood pressure; SIRS = systemic inflammatory response syndrome; SOFA = Sequential Organ Failure Assessment; SVO2= mixed venous oxygen
saturation; T = temperature; TBSA = total body surface area; UOP = urine output; VR = ventilatory rate; WBC = white blood cell count.
a
Either documented or suspected.
b
No discrete number of symptoms was published; the stated goal was to facilitate bedside diagnosis instead of creating standardized diagnostic criteria [19].
c
Defined as SBP <90 mm Hg or reduction in SBP ‡40 mm Hg.
 DETECTING BURN INFECTION AND SEPSIS 5

(Fig. 1) [23]. The following are notable examples of candi-

date serum biomarkers.
PCT (ng/mL), suPAR (ng/mL)

10
and Lactate (mmol/L)

CRP Procalcitonin. Procalcitonin (PCT) is a serum protein with

8
considerable potential for use in the detection of sepsis in pa-
6 tients with severe burn injuries. Measurement of PCT has
PCT
Lactate gained attention in the critical care setting because of its notably
4
consistent correlation with the course of bacterial infection. It is
2
an endocrine hormone of the thyroid gland and precursor to
IL-6
TNF- calcitonin that plays an indirect role in calcium metabolism and
IL-10
0 has undetectable serum concentrations under normal condi-
0 1 2 6 12 24 48 72
Time (hours) tions. However, during a systemic inflammatory response and
in the presence of microbial endotoxin, it is released into the
FIG. 1. Time course of sepsis biomarker elevation after circulation by non-endocrine tissues including the lung, kidney,
pathogen exposure. Borrowed and modified with permis- and omentum [24]. Unlike other serum biomarkers, the utility
sion from University of California Davis Department of Pa- of PCT lies in its responsiveness to both inflammatory medi-
thology [23]. CRP = C-reactive protein, IL-6 = interleukin 6, ators and bacterial lipopolysaccharide (LPS) [24,25]. Similarly,
IL-10 = interleukin 10, PCT = procalcitonin, TNF-a = tissue fungal toxins elicit an increase in PCT, but to a lesser extent
necrosis factor alpha.
than bacterial LPS because of activation of an alternative cy-
Downloaded by Universiteit Utrecht from www.liebertpub.com at 10/08/20. For personal use only.

tokine response involving interferon (IFN)-g [26]. Although

Sepsis-3, the ISBI practice guidelines published in 2018 af- PCT is useful in the detection of bacterial and fungal infections,
terward re-affirmed the previously instituted ABA consensus its measurement is limited in detecting viruses, likely because
definitions as standard of care in burn patients [8]. Ironically, a of the highly potent viral inhibition of tumor necrosis factor
prospective study by Yan et al. [22] published the same year (TNF)-a by IFN-g, effectively suppressing measurable PCT.
compared the predictive value of ABA criteria, the afore- With regard to bacterial sepsis, a meta-analysis in 2019 by
mentioned Mann-Salinas criteria, and Sepsis-3 in adult burn Kondo et al. [27] showed an AUROC of 0.84 with a pooled
patients, discovering that they correctly predicted sepsis in sensitivity of 80% and specificity of 75% (Table 3). Although
59%, 28%, and 85% of patients, respectively, at 48 hours prior a sizable proportion of the literature acknowledges PCT’s
to clinical diagnosis of sepsis. Although at this time the evi- diagnostic accuracy in early detection of bacterial sepsis in
dence most supports Sepsis-3 for the reliable diagnosis of burn the critically ill, its clinical utility has been called into
sepsis, this has not been formally embraced in treatment question. Notably, Jensen et al. [28] in 2011 published the
guidelines and remains to be corroborated. results of the Procalcitonin and Survival Study (PASS) ran-
domized clinical trial, which recruited 1,200 critically ill
patients to evaluate the effectiveness of PCT-guided early
Serum biomarkers investigated for early
antibiotic therapy in improving sepsis survival. This strategy
detection of sepsis
yielded an absolute risk reduction of 0.6% contrasted with the
The lack of robust clinical criteria capable of timely and standard of care, which was not statistically significant.
specific diagnosis of burn sepsis as well as the magnitude of Nonetheless, mounting evidence exists for the diagnostic
clinical impact imparted by early treatment has triggered a value of PCT in the care of burned patients. The relation
massive undertaking to search for methods for the early de- between PCT and sepsis has been explored throughout the
tection and diagnosis of sepsis. A wide spectrum of inflam- past three decades, and studies continue to report evidence in
matory and molecular serum biomarkers identifiable early in favor of PCT’s diagnostic utility in burn patients. One such
the course of infection has already undergone investigation investigation assessed the diagnostic capabilities of PCT

Table 3. Accuracy of Serum Biomarkers for Detection of Sepsis in Burn Injury

Studied
Sensitivity Specificity exclusively in
Marker (%) (%) AUROC burn patients Predictive value Type of study
PCT 80 75 0.84 N Sepsis in ICU Meta-analysis [27]
74 88 0.92 Y Early sepsis Meta-analysis [30]
77 65 0.83 Y Early sepsis Meta-analysis [31]
CRP 80 61 0.73 N Early sepsis Meta-analysis [36]
92 58 0.82 Y Sepsis Prospective [54]
IL-6 68 73 0.80 N Sepsis Meta-analysis [38]
IL-8 75–80 75–80 0.88 Y Sepsis Prospective [42]
IL-10 85 84 0.87 Y Sepsis; death Prospective [55]
92 93 NR Y Sepsis; survival Observational [56]
TNF-a 68 88 0.93 N Sepsis in neonatal infants Meta-analysis [56]
mRNA (RT-PCR) 91 80 NR N Sepsis in ICU Observational [53]
CRP = C-reactive protein; IL = interleukin; mRNA = messenger RNA; N = no; NR = not reported; PCT = procalcitonin; RT-PCR = reverse
transcriptase polymerase chain reaction; TNF-a = tumor necrosis factor; Y = yes.
 6 TORRES ET AL.

compared with C-reactive protein (CRP), erythrocyte sedi- Interleukin-10 mediates the immune response to infection,
mentation rate (ESR), and white blood cell (WBC) count. Not primarily inhibiting inflammatory activities of macrophages
only was PCT found to have excellent diagnostic accuracy in and Th1, NK, and B cells in an attempt to downregulate the
the prediction of infection, but it also was the only serum marker pro-inflammatory state [43]. Assessment of IL-10 in isolation
that displayed an observable difference between septic and from other serum biomarkers has limited utility, yet it may
non-septic cases [29]. Two recently completed meta-analyses provide further insight into the prognostic trajectory of burn
demonstrated findings in burn sepsis largely consistent with injury and sepsis. Similar to findings in IL-8, burn patients
each other and earlier works, with an AUROC of 0.92 [30] and who develop sepsis tend to have remarkably higher concen-
0.83 [31]. Furthermore, a later study determined that elevated trations of IL-10, which remain elevated for as long as 14 days
serial measurements of PCT during the first five days after burn after injury and into the development of infection [44,45].
injury were strongly associated with the development of sepsis
[32]. In addition to its diagnostic ability, contrasting concen- Tumor Necrosis Factor-a. During the acute period after a
trations of PCT may be able to help distinguish gram-negative severe burn, macrophages on location participate in the immune
from gram-positive bacterial infection, with significantly higher response by secreting tumor necrosis factor (TNF)-a which
concentrations indicating gram-negative infections [33]. upregulates the activity of pro-inflammatory cytokines [46].
Because TNF-a also is released after infection, studies have
C-Reactive Protein. C-reactive protein (CRP) is a thor- explored its value in the recognition of sepsis. Significantly
oughly studied acute-phase reactant and serum biomarker of elevated concentrations of TNF-a were measured in burned
inflammation that is produced primarily in the liver and up-
Downloaded by Universiteit Utrecht from www.liebertpub.com at 10/08/20. For personal use only.

patients at the time of admission, with a second elevation of

regulated by interleukin-6 (IL-6) [34]. Its concentration cor- TNF-a occurring later in the hospital course [44,46]. Subsequent
relates strongly with burn size, and higher concentrations tend studies reported that lower ratios of TNF-a to IL-10 were highly
to be found in women and non-survivors [35]. Beyond these predictive of repeated episodes of infection and were correlated
findings, its utility for the detection of sepsis is not supported, inversely with burn size [47]. However, the utility of TNFa in
as demonstrated in a meta-analysis of nine studies published the detection of sepsis still is limited because of conflicting data
in 2018 by Tan et al. [36], reporting a sensitivity of 0.80 and in some populations. For example, TNF-a concentrations do not
specificity of 0.61. Jeschke et al. [35] carried out the largest become elevated in elderly patients regardless of sepsis devel-
cohort study to date specifically for sepsis detection in burned opment during the first 30 days after the burn, which may be
patients, with a total of 918 pediatric enrollees included in the related to a latent or impaired host immune response [48].
analysis. The investigation confirmed that CRP was unable to
predict sepsis at any change in threshold between 0 and 20 mg/ Molecular Markers. In recent years, emerging interest has
dL. Although widely available, its use is not recommended in developed for cellular and molecular laboratory techniques in
routine burn critical care. the detection of sepsis in burned patients based on encour-
aging findings in the fields of genomics and transcriptomics.
Interleukins 6, 8, and 10. Interleukin (IL)-6 is a major By using widely available genomics assays to identify sets of
acute-phase reactant released by T cells and macrophages at the relevant genetic markers, these tools may provide highly
site of injury, and its utility in the diagnosis of infection has accurate and reliable means of early recognition of sepsis. For
been well studied [37,38]. In a meta-analysis of 20 studies on example, carriers of particular alleles relating to expression
adults with SIRS, IL-6 was determined to be of diagnostic value of interleukins, TNF-a, and Toll-like receptor 4 have a higher
comparable to that of PCT in its ability to detect sepsis, with IL- risk of sepsis and death after severe burns [49–51].
6 having a pooled sensitivity of 0.68 and specificity of 0.73 Epigenetic processes, such as DNA methylation and histone
[38]. In burned patients, patterns of elevated concentrations are modification, may also be useful in early detection of sepsis and
biphasic, peaking at both early and late time points after severe can specify the bacterial etiology [52]. Studies on transcribed
burns. Monitoring for persistently high concentrations of IL-6 mRNA and smaller, non-coding microRNA that are involved in
may raise suspicion of upcoming sepsis in patients with burns the pro-inflammatory state indicate the potential of a diagnostic
[39]. However, IL-6 remains questionable in its clinical utility pattern for identifying sepsis. One study measured expression
for the detection of sepsis. Tsalik et al. [40] studied a cohort of of a panel of seven cytokine genes using real-time reverse
336 adults with suspected sepsis, and IL-6 was demonstrated to transcriptase polymerase chain reaction together with non-
have an insignificant value in discriminating between severity linear neural network analysis and demonstrated that they
of infection compared with PCT and CRP. could predict sepsis accurately as early as four days prior to
Interleukin-8 is a pro-inflammatory cytokine that is notably onset with a sensitivity and specificity of 91.4% and 80.2%,
released during the acute period of injury and acts as a potent respectively [53]. Proteomics and metabolomics are two
attractant for neutrophils, promoting their migration into the additional fields of analysis where pattern recognition and
affected area [41]. Investigations have highlighted a clinical molecular expression can be utilized for sepsis recognition.
value in IL-8 that is primarily prognostic rather than diagnostic,
in which a relation exists between its serum concentration and
Conclusions
the likelihood of sepsis and survival. In a large cohort study by
Kraft et al., IL-8 was measured in 468 pediatric burn patients Early intervention is critical to improve outcomes of sep-
with TBSA of >30%, in which concentrations above a threshold sis, but early treatment necessitates early detection. Despite
of 234 pg/mL indicated a higher likelihood of multiorgan dys- the existence of criteria specific for burn sepsis published by
function, sepsis, and death [42]. Persistently elevated IL-8 cor- the American Burn Association, they have not been reliable
relates with a poorer prognosis and high possibility of infection in prospective studies beyond the contemporary Sepsis-3
in burn patients [39]. criteria. Investigation of serum biomarkers of sepsis also has
 DETECTING BURN INFECTION AND SEPSIS 7

yet to produce a gold standard or widely implemented assay. 16. Pereira CT, Murphy KD, Herndon DN. Altering metabo-
Hope persists that reliable criteria and assays will be identi- lism. J Burn Care Rehabil 2005;26:194–199.
fied for the early detection of burn sepsis. Meanwhile, a 17. Bone RC, Balk RA, Cerra FB, et al. Definitions for sepsis
thorough understanding and appropriate application of the and organ failure and guidelines for the use of innovative
available diagnostic criteria and assays are paramount in therapies in sepsis. Chest 1992;101:1644–1655.
providing care of severe burns. 18. Tridente A. Sepsis 3 and the burns patient: Do we need
Sepsis 3.1? Scars Burn Heal 2018;4:205951311879065.
19. Levy MM, Fink MP, Marshall JC, et al. 2001 SCCM/
Acknowledgments
ESICM/ACCP/ATS/SIS International Sepsis Definitions
We thank Dr. Nam K. Tran and the University of Cali- Conference. Crit Care Med 2003;31:1250–1256.
fornia Davis Department of Pathology for allowing us to 20. Greenhalgh DG, Saffle JR, Holmes JH, et al. American
reproduce a modified version of Figure 1 [23]. Burn Association Consensus Conference to Define Sepsis
and Infection in Burns The American Burn Association
Author Disclosure Statement Consensus Conference on Burn Sepsis and Infection Group
2007;776–790.
The authors have no disclosures related to this article. 21. Mann-Salinas EA, Baun MM, Meininger JC, et al. Novel
predictors of sepsis outperform the American Burn Asso-
References ciation sepsis criteria in the burn intensive care unit patient.
J Burn Care Res 2013;34:31–43.
1. Greenhalgh DG. Sepsis in the burn patient: A different prob-
Downloaded by Universiteit Utrecht from www.liebertpub.com at 10/08/20. For personal use only.

22. Yan J, Hill WF, Rehou S, et al. Sepsis criteria versus

lem than sepsis in the general population. Burn Trauma 2017; clinical diagnosis of sepsis in burn patients: A validation of
5:1–10. current sepsis scores. Surgery 2018;164:1241–1245.
2. Data & Reports j Sepsis j CDC [Internet]. Available at 23. Laboratory Best Practice j UC Davis Department of Pa-
https://www.cdc.gov/sepsis/datareports/index.html (Last ac- thology and Laboratory Medicine [Internet]. Available at:
cessed May 25, 2020). https://health.ucdavis.edu/pathology/lab-best-practice/sepsis
3. Williams FN, Herndon DN, Hawkins HK, et al. The leading .html (Last accessed August 13, 2020).
causes of death after burn injury in a single pediatric burn 24. Müller B, White JC, Nylén ES, et al. Ubiquitous expression
center. Crit Care 2009;13:1–7. of the calcitonin-I gene in multiple tissues in response to
4. Bloemsma GC, Dokter J, Boxma H, et al. Mortality and sepsis. J Clin Endocrinol Metab 2001;86:396–404.
causes of death in a burn centre. Burns 2008;34:1103–1107. 25. Linscheid P, Seboek D, Nylen ES, et al. In vitro and in vivo
5. Lachiewicz AM, Hauck CG, Weber DJ, et al. Bacterial calcitonin I gene expression in parenchymal cells: A novel
infections after burn injuries: Impact of multidrug resis- product of human adipose tissue. Endocrinology 2003;144:
tance. Clin Infect Dis 2017;65:2130–2136. 5578–5584.
6. Huzar TF, Cross JM. Ventilator-associated pneumonia in 26. Dou Y, Du J, Liu H, et al. The role of procalcitonin in the
burn patients: A cause or consequence of critical illness? identification of invasive fungal infection—A systemic re-
Expert Rev Respir Med 2011;5:663–673. view and meta-analysis. Elsevier.
7. Carr JA, Phillips BD, Bowling WM. The utility of bron- 27. Kondo Y, Umemura Y, Hayashida K, et al. Diagnostic
choscopy after inhalation injury complicated by pneumonia value of procalcitonin and presepsin for sepsis in critically
in burn patients: Results from the National Burn Re- ill adult patients: A systematic review and meta-analysis. J
pository. J Burn Care Res 2009;30:967–974. Intensive Care 2019;7:1–13.
8. Allorto N, Atieh B, Bolgiani A, et al. ISBI Practice 28. Jensen JU, Hein L, Lundgren B, et al. Procalcitonin-guided
Guidelines for Burn Care 2. Burns 2018;44:1617–1706. interventions against infections to increase early appropriate
9. Fàbregas N, Ewig S, Torres A, et al. Clinical diagnosis of antibiotics and improve survival in the intensive care unit: A
ventilator associated pneumonia revisited: Comparative randomized trial. Crit Care Med 2011;39:2048–2058.
validation using immediate post-mortem lung biopsies. 29. Barati M, Alinejad F, Bahar MA, et al. Comparison of
Thorax 1999;54:867–873. WBC, ESR, CRP and PCT serum levels in septic and non-
10. Mosier MJ, Pham TN. American Burn Association prac- septic burn cases. Burns 2008;34:770–774.
tice guidelines for prevention, diagnosis, and treatment 30. Ren H, Li Y, Han C, et al. Serum procalcitonin as a di-
of ventilator-associated pneumonia (VAP) in burn patients. agnostic biomarker for sepsis in burned patients: A meta-
J Burn Care Res 2009;30:910–928. analysis. Burns 2015;41:502–509.
11. Church D, Elsayed S, Reid O, et al. Burn wound infections. 31. Cabral L, Afreixo V, Almeida L, et al. The use of pro-
Clin Microbiol Rev 2006;19:403–434. calcitonin (PCT) for diagnosis of sepsis in burn patients: A
12. Mayhall CG. The epidemiology of burn wound infections: meta-analysis. PLoS One 2016;11(12):1–16.
Then and now. Clin Infect Dis 2003;37:543–550. 32. Cabral L, Afreixo V, Meireles R, et al. Procalcitonin ki-
13. Woolfrey BF, Fox JM, Quall CO. An evaluation of burn netics after burn injury and burn surgery in septic and non-
wound quantitative microbiology I: Quantitative eschar septic patients: A retrospective observational study. BMC
cultures. Am J Clin Pathol 1981;75:532–537. Anesthesiol 2018;18:1–10.
14. Nunez Lopez O, Cambiaso-Daniel J, Branski LK, et al. 33. Cabral L, Afreixo V, Meireles R, et al. Evaluation of pro-
Therapeutics and clinical risk management dovepress pre- calcitonin accuracy for the distinction between gram-
dicting and managing sepsis in burn patients: Current per- negative and gram-positive bacterial sepsis in burn patients.
spectives. Ther Clin Risk Manag 2017;13:1107. J Burn Care Res 2019;40:112–119.
15. Singer M, Deutschman CS, Seymour C, et al. The Third 34. Gabay C, Kushner I. Acute-phase proteins and other sys-
International Consensus Definitions for Sepsis and Septic temic responses to inflammation. N Engl J Med 1999;340:
Shock (Sepsis-3). JAMA 2016;315:801–810. 448–454.
 8 TORRES ET AL.

35. Jeschke MG, Finnerty CC, Kulp GA, et al. Can we use dictor of increased susceptibility to infections. Front Public
C-reactive protein levels to predict severe infection or sepsis Heal 2016;4(Oct):1–11.
in severely burned patients? Int J Burns Trauma 2013;3:137– 48. Stanojcic M, Vinaik R, Jeschke MG. Status and challenges
143. of predicting and diagnosing sepsis in burn patients. Surg
36. Tan M, Lu Y, Jiang H, et al. The diagnostic accuracy of Infect 2018;19:168–175.
procalcitonin and C-reactive protein for sepsis: A system- 49. Accardo Palumbo A, Forte GI, Pileri D, et al. Analysis of
atic review and meta-analysis. J Cell Biochem 2019;120: IL-6, IL-10 and IL-17 genetic polymorphisms as risk fac-
5852–5859. tors for sepsis development in burned patients. Burns 2012;
37. Scheller J, Chalaris A, Schmidt-Arras D, et al. The pro- and 38:208–213.
anti-inflammatory properties of the cytokine interleukin-6. 50. Shalhub S, Pham TN, Gibran NS, et al. Tumor necrosis
Biochim Biophys Acta Mol Cell Res 2011;1813:878–888. factor gene variation and the bisk of mortality after burn
38. Ma L, Zhang H, Yin YL, et al. Role of interleukin-6 to injury: A cohort study. J Burn Care Res 2009;30:105–111.
differentiate sepsis from non-infectious systemic inflam- 51. Barber RC, Chang LYE, Arnoldo BD, et al. Innate im-
matory response syndrome. Cytokine 2016;88:126–135. munity SNPs are associated with risk for severe sepsis after
39. Jeschke MG, Gauglitz GG, Finnerty CC, et al. Survivors burn injury. Clin Med Res 2006;4:250–255.
versus nonsurvivors postburn: Differences in inflammatory 52. Tendl KA, Schulz SMF, Mechtler TP, et al. DNA meth-
and hypermetabolic trajectories. Ann Surg 2014;259:814– ylation pattern of CALCA in preterm neonates with bac-
823. terial sepsis as a putative epigenetic biomarker. Epigenetics
40. Tsalik EL, Jaggers LB, Glickman SW, et al. Discriminative 2013;8:1261–1267.
Downloaded by Universiteit Utrecht from www.liebertpub.com at 10/08/20. For personal use only.

value of inflammatory biomarkers for suspected sepsis. J 53. Lukaszewski RA, Yates AM, Jackson MC, et al. Pre-
Emerg Med 2012;43:97–106. symptomatic prediction of sepsis in intensive care unit
41. Gauglitz GG, Finnerty CC, Herndon DN, et al. Are serum patients. Clin Vaccine Immunol 2008;15:1089–1094.
cytokines early predictors for the outcome of burn patients 54. Cxakir Madenci Ö, Yakupoǧlu S, Benzonana N, et al. Eva-
with inhalation injuries who do not survive? Crit Care luation of soluble CD14 subtype (presepsin) in burn sepsis.
2008;12:1–8. Burns 2014;40:664–669.
42. Kraft R, Herndon DN, Finnerty CC, et al. Predictive value 55. Csontos C, Foldi V, Pálinkas L, et al. Time course of pro-
of IL-8 for sepsis and severe infections after burn injury: A and anti-inflammatory cytokine levels in patients with
clinical study. Shock 2015;43:222–227. burns: Prognostic value of interleukin-10. Burns 2010;36:
43. Asadullah K, Sterry W, Volk HD. Interleukin-10 therapy: 483–494.
Review of a new approach. Pharmacol Rev 2003;55:241– 56. Lv B, Huang J, Yuan H, et al. Tumor necrosis factor-a as a
269. diagnostic marker for neonatal sepsis: A meta-analysis. Sci
44. Finnerty CC, Herndon DN, Chinkes DL, et al. Serum cy- World J 2014;2014.
tokine differences in severely burned children with and
without sepsis. Shock 2007;27:4–9. Address correspondence to:
45. Pileri D, Accardo Palombo A, D’Amelio L, et al. Con- Dr. Joshua M. Peterson
centrations of cytokines IL-6 and IL-10 in plasma of burn Division of Plastic Surgery
patients: Their relationship to sepsis and outcome. Ann Department of Surgery
Burns Fire Disasters 2008;21:182–185. University of Texas Medical Branch
46. Abdel-Hafez NM, Saleh Hassan Y, El-Metwally TH. A 301 University Boulevard
study on biomarkers, cytokines, and growth factors in 6.124 McCullough Building
children with burn injuries. Ann Burns Fire Disasters 2007; Galveston, TX 77555-0724
20:89–100. USA
47. Tsurumi A, Que YA, Ryan CM, et al. TNF-a/IL-10 ratio
correlates with burn severity and may serve as a risk pre- E-mail: [email protected]