Skip Navigation LinksHome > July/August 2014 - Volume 35 - Issue 4 > Predictors of Serious Bacterial Infections in Pediatric Burn...
Journal of Burn Care & Research:
doi: 10.1097/BCR.0b013e3182a23211
Original Articles

Predictors of Serious Bacterial Infections in Pediatric Burn Patients With Fever

Vyles, David DO*; Sinha, Madhumita MD*; Rosenberg, David I. MD*; Foster, Kevin N. MD, MBA; Tran, Melissa; Drachman, David PhD

Free Access
Article Outline
Collapse Box

Author Information

From the *Department of Pediatrics, Arizona Children’s Center, Department of Surgery, Arizona Burn Center, and Department of Research, Maricopa Integrated Health System, Phoenix.

This study was presented at the Pediatric Academic Societies Annual Meeting, at Washington, DC, in May 2013.

Address correspondence to Madhumita Sinha, MD, Department of Pediatrics, Maricopa Integrated Health System, 2601 E. Roosevelt St., Phoenix, Arizona 85008.

Collapse Box


To determine predictors of serious bacterial infections in pediatric burn patients with fever (core temp ≥38.5°C), the authors conducted a retrospective review of medical records of pediatric (0–18 years) patients admitted to the Arizona Burn Center between 2008 and 2011 with greater than 5% TBSA and inpatient hospitalization for ≥72 hours. The study group comprised patients with a febrile episode during their inpatient stay. Serious bacterial infection (the primary outcome variable) was defined as: bacteremia, urinary tract infection, meningitis (blood, urine, or cerebrospinal fluid culture positive for a pathogen respectively), pneumonia, line, and wound infection. A generalized estimating equation analysis was done to predict the presence or absence of serious bacterial infection. Of 1082 pediatric burn patients hospitalized during the study period, 353 met the study eligibility criteria. A total of 108 patients (30.6%) had at least one fever episode (fever group). No difference in demographic characteristics was noted between the fever and no-fever groups; significant differences were observed for: third-degree TBSA, second-degree TBSA, total operating room visits, length of stay, Injury Severity Score, and death. A total of 47.2% of the patients had one or more episodes of fever with serious bacterial infection. In a generalized estimating equation predictive model, presence of a central line, second-, and third-degree TBSA were predictive of serious bacterial infection in burn patients with fever. In this study, individual clinical variables such as tachypnea and tachycardia were not predictive of serious bacterial infections, but the presence of a central line, and larger TBSA were significant predictors of serious bacterial infections. Younger age (P =.08) and ventilator support (P =.057) also approached significance as predictors of serious bacterial infections.

Sepsis continues to be one of the leading causes of morbidity and mortality among pediatric burn patients.1 Early recognition of presence of infection and prompt institution of appropriate therapy is likely to prevent rapid clinical deterioration and improve clinical outcomes in children with significant burn injury.

The modified systemic inflammatory response syndrome (SIRS) in children is defined as the presence of two or more of the following criteria: abnormal core body temperature, tachycardia, tachypnea, and abnormal leucocyte count.2 Sepsis is SIRS in the presence of or as a result of proven or suspected infection.3 Because burn injury is characterized by an acute inflammatory response, mere increase of temperature with coexistent tachypnea and tachycardia does not necessarily herald the presence of an underlying infection. Hence, diagnosis of serious bacterial infections or predicting onset of sepsis in a febrile pediatric burn patient with an already accentuated hypermetabolic state remains a challenging issue for physicians. The American Burn Association (ABA) in its Consensus Conference in 2007 reiterated that a SIRS-like state is inherently associated with burn injury, and the need for more specific criteria for early detection of infection in burn patients was recognized and formulated.4 The modified criteria published by ABA for pediatric sepsis include: abnormality of temperature (>39°C or <36.5°C), presence of tachycardia (>2 SD above norms for age), tachypnea (>2 SD above norms for age), thrombocytopenia (<2 SD below norms for age), hyperglycemia, and enteral feed intolerance.4 As per ABA recommendations, presence of more than three of these criteria should be considered highly suspicious and should trigger concern for an underlying infection. However, recent studies have suggested that the ABA’s trigger criteria for infection and sepsis in burn patients do not correlate well with bacteremia.5,6 These studies, however, have primarily focused on adult populations and their generalizability to pediatric age groups has not been well documented. The objective of our current study was to determine predictors of serious bacterial infections in pediatric burn patients with fever (core temp ≥38.5°C). This temperature threshold was recommended at the International Pediatric Sepsis Conference for definition of sepsis and organ dysfunction; it was believed to improve specificity and reflect current intensive care unit practice.2 Serious bacterial infection is defined as a positive microbiological culture from a sterile sample such as blood culture positive for a pathogen (bacteremia), urine culture positive for a pathogen (urinary tract infection), cerebrospinal fluid (CSF) culture positive for a pathogen (meningitis), pneumonia (chest radiograph findings positive for a lobar infiltrate), line infection (catheter-associated blood stream infection), and wound infection (positive wound culture for a pathogen).

Back to Top | Article Outline



We conducted a retrospective review of medical records of pediatric patients between birth and 18 years of age who were admitted to the hospital with a burn-related injury. Eligible patients were identified from the Arizona Burn Center burn registry database of all pediatric hospitalized burn patients from 2008 to 2011; additional demographic and clinical data were obtained through review of medical records of individual patients.

Back to Top | Article Outline

This study was conducted in the Arizona Burn Center, a stand-alone center of excellence with a 37-bed facility within a large tertiary care medical center. The Arizona Burn Center provides care to approximately 1000 inpatients and 5500 adult and pediatric outpatients annually. The Institutional Review Board of the hospital approved the study.

Back to Top | Article Outline

Because infection and onset of sepsis are late complications in burns,1 pediatric burn patients hospitalized for 72 hours or more, with greater than 5% TBSA involved and who developed a fever (core temperature of 38.5°C or higher), were eligible for inclusion in the study. Data points available from the burn registry included: patient’s age, sex, race/ethnicity, injury site, injury cause, TBSA, Injury Severity Score (ISS) scores, total operative procedures and operating room (OR) visits, total hospital days, intensive care unit (ICU) days, ventilator days, and discharge status. Additional information obtained from review of individual patient medical records included: clinical variables: total fever days; number of fever episodes 48 hours apart; individual clinical components of ABA criteria during febrile episodes, including presence of tachypnea (age specific) and tachycardia (age specific); these were entered in the multivariate analysis. The information also included laboratory variables: an abnormal leucocyte count (>15,000 or <5,000 μl); blood, urine, CSF, and wound culture reports; as well as reports of chest radiograph findings. The primary outcome variable was presence or absence of serious bacterial infection. Each fever episode (48 hours apart) during hospitalization was considered as a “sentinel event” and analyzed separately.

Back to Top | Article Outline
Statistical Analysis

Statistical analysis was conducted using SPSS for Windows 20.0 (SPSS Inc., Chicago, IL). We used descriptive statistics to describe sample characteristics and t-tests or χ2 tests to compare demographic and clinical characteristics between patients with and without fever. A P value ≤.05 was considered significant. A generalized estimating equations (GEEs) were used for predictive modeling. The GEE model is used for analysis of correlated data and it enabled us to study variables occurring for each fever episode (such as tachypnea) and variables occurring only once for each patient (such as sex).

Back to Top | Article Outline


A total of 1082 children with burn injuries were hospitalized during the 4-year (2008–2011) study period. Of these, 353 fulfilled study eligibility criteria and were included. A total of 108 patients (30.6%) had at least one fever episode during their inpatient stay. The rest (245 patients) did not have fever during the course of their hospitalization. There were no significant differences in demographic characteristics (age, sex, race/ethnicity) between burn patients who had fever and those who remained afebrile during their hospitalization (Table 1). However, febrile burn patients had higher TBSA involved, greater number of hospital days, greater number of OR visits, and procedures, higher ISS, and death compared with children with no fever (Table 1). There was no significant difference between the fever and no-fever groups relating to length of ICU stay (P value was.07 according to the Mann–Whitney U test comparing medians), and for ventilator days (P value was.92 according to the Mann–Whitney U test comparing medians). Of the 106 febrile patients for whom complete data were available, 49 or 46.2% had one or more episodes of fever with serious bacterial infection. Thirty-eight patients or 35.2% had a single episode of fever with serious bacterial infection, 10 patients or 9.3% had exactly two episodes of fever with serious bacterial infection, and one patient or 0.9% had exactly three episodes of fever with associated serious bacterial infection.

Table 1.
Table 1.
Image Tools

Table 2 delineates the pathogens isolated from blood, urine, and wound cultures during individual febrile episodes among children with burns at our facility. There were no patients with a pathogen isolated from CSF in this cohort. All patients with blood/urine culture–proven infections were treated with antibiotics, patients with Candida infections in blood or urine were treated with fluconazole.

Table 2.
Table 2.
Image Tools

There were a total of 14 patients with chest radiograph findings that proved pneumonia during their first febrile episode. Of these, 12 patients were treated with intravenous antibiotics. Two patients were observed without antibiotics and showed clinical improvement, one of them tested positive for respiratory syncytial virus. Four patients had chest radiograph findings consistent with pneumonia during their second febrile episode; three of them were intubated and received mechanical ventilation, bacterial pathogens were isolated from their tracheal aspirate and they were treated with appropriate parenteral antibiotics. One patient had a tracheostomy and was being simultaneously treated for methicillin-resistant Staphylococcus aureus and Klebsiella urosepsis. One patient with chest radiograph–proven ventilator-associated pneumonia was diagnosed and treated during their third febrile episode, cultures were positive for Acinetobacter spp.

A GEE analysis of the data was done with presence/absence of serious bacterial infection as the outcome variable. The predictor variables used in the GEE included: sex, age, presence of tachycardia, tachypnea during febrile episodes, ISS, second- and third-degree TBSA, total hospital days, need for ICU stay, need for ventilator support, central line, indwelling Foley catheter, total operative procedures, and number of OR visits (Table 3). The fever criterion, a core body temperature >38.5°C, was positive for every case for which it was recorded, therefore it had to be excluded from the model. A total of 128 cases were included in the model. The presence of a central line and larger TBSA involvement with secondand third-degree burns were significant predictors of serious bacterial infection (Table 3).

Table 3.
Table 3.
Image Tools
Back to Top | Article Outline


The clinical significance of fever in a child with burns is a complex issue. Distinguishing an infectious cause from an inflammatory process and implementation of appropriate management strategies often proves to be a diagnostic and therapeutic challenge. In this study we looked at predictors of serious bacterial infections in a cohort of children who had at least one febrile episode during their inpatient stay. Our multivariate analysis with several key demographic and clinical variables also included some individual components of ABA’s pediatric sepsis clinical criteria such as presence of tachypnea, and tachycardia. These clinical criteria, however, were not found to be significant predictors of serious bacterial infection because they are nonspecific among children and are a frequent manifestation in a range of disease processes.2 However, larger TBSA (second- and third-degree burns) and presence of central lines were significant predictors of serious bacterial infection in our study. In a recently published study investigating burn size and survivability of pediatric burn patients using current treatment regimens, a burn size of greater than 60% was found to be a crucial threshold influencing morbidity and mortality; 9% of children in this study had sepsis with a significant increase in incidence of sepsis among children with a larger TBSA involved.7 A study by Rodgers et al8 also found that pediatric burn patients with flame and inhalation injury, TBSA burned >30%, and full-thickness burns were at a higher risk of having infectious complications. Interestingly, in contrast to the Rodgers et al8 study, wound infections were infrequent in our study, however, catheter-associated risk of serious bacterial infection was a significant predictor in both studies.

As with other retrospective studies, our study has certain limitations, including a lack of uniformity in documentation, which resulted in missing data in certain cases. Also, there was variability in laboratory testing so certain lab data such as that regarding prealbumin and transferrin could not be included as predictor variables because they were not available in all cases. However, documentation was adequate for the major clinical and lab variables so we feel that missing data are unlikely to have affected the final study conclusions.

Fever in burns, especially in the pediatric age group, remains a difficult therapeutic challenge. In our study we found individual clinical components of the ABA’s trigger criteria for infection and sepsis, such as presence of tachypnea and tachycardia, not to be helpful in identifying children with serious bacterial infection. Larger prospective multicentric studies are needed to identify specific risk factors including significance of quantification of serum inflammatory mediators to aid in early diagnosis of underlying infections and initiation of appropriate antimicrobial treatment to prevent complications and improve outcomes.

Back to Top | Article Outline


1. Sheridan RL. Sepsis in pediatric burn patients. Pediatr Crit Care Med. 2005; 6:(Suppl 3)S112–9

2. Goldstein B, Giroir B, Randolph A. International Consensus Conference on Pediatric Sepsis International pediatric sepsis consensus conference: definitions for sepsis and organ dysfunction in pediatrics. Pediatr Crit Care Med. 2005; 6:2–8

3. Bone RC, Sprung CL, Sibbald WJ. Definitions for sepsis and organ failure. Crit Care Med. 1992; 20:724–6

4. Greenhalgh DG, Saffle JR, Holmes JH 4th, et al. American Burn Association Consensus Conference on Burn Sepsis and Infection Group American Burn Association consensus conference to define sepsis and infection in burns. J Burn Care Res. 2007; 28:776–90

5. Hogan BK, Wolf SE, Hospenthal DR, et al. Correlation of American Burn Association sepsis criteria with the presence of bacteremia in burned patients admitted to the intensive care unit. J Burn Care Res. 2012; 33:371–8

6. Mann-Salinas EA, Baun MM, Meininger JC, et al. Novel predictors of sepsis outperform the American Burn Association sepsis criteria in the burn intensive care unit patient. J Burn Care Res. 2013; 34:31–43

7. Kraft R, Herndon DN, Al-Mousawi AM, Williams FN, Finnerty CC, Jeschke MG. Burn size and survival probability in paediatric patients in modern burn care: a prospective observational cohort study. Lancet. 2012; 379:1013–21

8. Rodgers GL, Mortensen J, Fisher MC, Lo A, Cresswell A, Long SS. Predictors of infectious complications after burn injuries in children. Pediatr Infect Dis J. 2000; 19:(10)990–5

Copyright © 2014 by the American Burn Association


Article Tools