PREDICTION TOOL

Criteria abstracted from The Users' Guides to the Medical Literature series in JAMA

Crit Care Med 2007 35(2):579-83.[abstract]

Review posted November 17, 2007

1. What is being studied?
   
   Objective: to determine the risk of sepsis and mortality after receiving blood products in severely burned children [30% total body surface area (TBSA)].
   
   Design: retrospective cohort
   
   
2. Are the results in the study valid?
   
   
   1. Was a representative group of patients completely followed up?
      
      Yes. Pediatric burn patients were enrolled if the TBSA was >30%; 277 patients were enrolled. Twenty-five patients that were septic on admission were excluded. All patients were followed from admission, to OR visits, or to time of onset of sepsis. Blood products were not leukoreduced and transfusion parameters were clinical signs of anemia and hemoglobin (Hgb) ≤8 mg/dL with target Hgb of 10-11 mg/dL. Overall, there were no differences between patients in terms of surgical burn treatment. All patients received the same topical medications. Inhalation injury was diagnosed by bronchoscopy. All patients were fed enterally. Patients who received Integra, an artificial dermis, were excluded from the study. Heimbach et al. showed that in adult burn patients the incidence of superficial, invasive and total infections were greater in patients with Integra® (1).
      
      
   2. Was follow-up sufficiently long and complete?
      
      The length of follow up was not explicitly stated. It is not stated what the length of follow up was, only from admission to onset of sepsis; not to discharge from the PICU or hospital, or if sepsis developed once transferred to the floor.
      
      
   3. Were all potential predictors included?
      
      The main predictor was blood product transfusions. Patients were further stratified by the percentage TBSA (<40%, 40-60%, and >60%) and presence or absence of inhalation injury. The cohort was divided into two groups: those patients who received >20 units of RBCs or high RBCs and those patients that received <20 units of RBCs or low RBCs. Covariates that could also impact the onset of sepsis included TBSA and presence or absence of inhalation injury.
      
      Inhalation injury was diagnosed by bronchoscopy upon admission or during the first OR visit (patients were taken to OR within 24 hours from admission as a standard of burn care at the authors institution). No confounding factors were included. However, there was no description of grades of inhalational injury. In addition, inhalation injury is an independent risk for sepsis. Rodgers, et al, found that all children in their study with a flame and inhalation injury developed an infection. They also found that pneumonia was one of the most common early infections. All patients with TBSA >30% also developed an infection (2). O Keefe, et al, found that the odds ratio for estimation death with inhalation injury was 3.4 (CI 1.9-6.0) (3).
      
      Presumably children both with larger burns and with inhalation injury were mechanically ventilated and had central venous catheters. Rodgers, et al, found that CVL infections were the most common cause of an infectious complication in children with burns (2). Mechanical ventilation alone is also a risk factor for sepsis. Ventilated patients and use of CVLs were not mentioned. Hospital trauma level was not mentioned and there were significantly more children with 40-60% and >60% TBSA burns versus children with <40% TBSA (sicker patients got septic).
      
      
   4. Did the investigators test the independent contribution of each predictor variable?
      
      Yes. They used multiple regression analysis. The authors state that an excellent predictor for sepsis was obtained with the three variables TBSA burn, presence of inhalational injury, and RBCs >20.
      
      
   5. Were outcome variables clearly and objectively defined?
      
      Sepsis criteria based on guidelines of the Society of Critical Care Medicine in conjunction with positive blood culture or the presence of organisms in the organs at autopsy. The criteria for burn sepsis are somewhat different from those published by the SCCM. The criteria listed in the article may be specific to burns but that is not stated. Delirium, refractory hypotension (SBP <90 mm Hg), and enteral feeding intolerance could be considered aspects of septic shock. A patient may receive more blood products based on the criteria in the article versus a patient who has sepsis with SIRS criteria and evidence of infection, thus, this association may have been overestimated based on the criteria used in the study.
      
      
3. What are the results?
   
   
   1. What is the prediction tool?
      
      Of 252 patients included in the study, 88 sustained inhalation injury, 61 developed sepsis and 30 died.
      
      The incidence of sepsis in children with TBSA 40-60% without inhalational injury was low in both the low and high RBC groups (0% and 6% respectively). In children with TBSA burns of 40-60% with inhalational injury the incidence of sepsis was higher (not statistically significant) in the high RBC group compared to the low RBC group (22% vs. 5%). The incidence of sepsis in children with TBSA 60% without inhalational injury was low in both the high and low RBC groups (5% vs. 9%). In children with >60% TBSA burn and inhalational injury the incidence of sepsis was higher (58% vs. 8%, p 0.05) in the high RBC group compared to the low RBC group. Children with large burns and who received FFP >5 units also had a higher risk of sepsis, 60% versus 7%.
      
      The univariate predictor, for all patients studied, that resulted in an increase in mortality was receiving RBCs >20 units with an odds ratio of 6.79. The mortality rate for the 60 patients that received more than 20 units of PRBC's was 30% versus the mortality rate of 6% for the 222 patients that received less than 20 units of PRBC's. In the patients with >60% TBSA burns, the odds ratio of death was 11.6 for those receiving >20 units of RBCs. However, confidence intervals were not stated.
      
      
   2. How well does the model categorize patients into different levels of risk?
      
      The study did not establish a dose-related causation between transfusion of RBCs and sepsis because there were only two categories studied: low RBC and high RBC groups with a cutoff between groups of 20 units.
      
      A multiple logistic regression equation was derived in order to independently model the roles of %TBSA, presence/absence of inhalation injury, and high or low RBC use.
      
      One area of concern in the current study is the amount of blood products received in the operating room. Anesthesiologists may transfuse to maintain a higher hematocrit than would be tolerated in the Intensive Care Unit.
      
      
4. Will the results help me in caring for my patients?
   
   Yes. We have similar patients in our PICU. In this study, there was an increased risk of sepsis in children with >60% TBSA and inhalational injury and who received >20 units of RBCs. Thus, limiting transfusions may be beneficial. However, with this study the amount of blood transfusions is not independent from the severity of illness and using a lower transfusion cutoff may not reduce transfusions in these patients.
   
   The study by Lacroix, et al, showing that a restrictive transfusion protocol (transfuse for Hgb of 7g/dL with goal of 8.5-9.5g/dL) in stable, critically ill pediatric patients decreased the number of transfusions. There was no difference in adverse outcomes (including nosocomial infections) between the restrictive or liberal-strategy groups (4). Although these data are not specific to burn patients, it is further evidence that restricting blood transfusions is not disadvantageous.
   
   Data specific to burn patients presented by Palmieri, et al, showed that when patients in a traditional transfusion group (transfuse to maintain Hgb ≥10 g/dL) were compared to a restrictive transfusion group (transfuse to maintain Hgb ≥7 g/dL) the patients in the traditional group received more blood. There were also more complications in the traditional group, but subgroup analysis of complications did not show any statistically significant differences. Finally, it costs more to transfuse to maintain Hgb at ≥10 g/dL (5).

References

1. Heimbach DM, Warden GD, Luterman A, Jordan MH, Ozobia N, Ryan CM, Voigt DW, Hickerson WL, Saffle JR, DeClement FA, Sheridan RL, Dimick AR. Multicenter Postapproval Clinical Trial of Integra Dermal Regeneration Template for Burn Treatment. J Burn Care Rehabil 2003; 24:42-48.
   
   
2. Rodgers GL, Mortensen J, Fisher MC, Lo A, Cresswell A, Long SS. Predictors of infectious complications after burn injuries in children. The Pediatric Infectious Disease Journal. 2000, Oct; 19(10):990-995.
   
   
3. O Keefe GE, Hunt JL, Purdue GF. An Evaluation of Risk Factors for Mortality after Burn Trauma and the Identification of Gender-Dependent Differences in Outcome. J Am Coll Surg. 2001; 192:153-160.
   
   
4. Lacroix J. Hebert PC. Hutchison JS. Hume HA. Tucci M. Ducruet T. Gauvin F. Collet JP. Toledano BJ. Robillard P. Joffe A. Biarent D. Meert K. Peters MJ. TRIPICU Investigators, Canadian Critical Care Trials Group, Pediatric Acute Lung Injury and Sepsis Investigators Network. Transfusion Strategies for Patients in Pediatric Intensive Care Units. New England Journal of Medicine. 2007, Apr 19; 356(16):1609-19.
   
   
5. Palmieri, TL, Lee, T, O Mara, MS, Greenhalgh, DG. Effects of a Restrictive Blood Transfusion Policy on Outcomes in Children With Burn Injury. J Burn Care Res. 2007; 28:65-70.

Last Updated: November 17, 2007

Fill out our Feedback Form | Editorial Notes | About Us