JBUR-4340; No. of Pages 10 burns xxx (2014) xxx–xxx

Available online at www.sciencedirect.com

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

Characterization of critically ill adult burn patients admitted to a Brazilian intensive care unit Edvaldo Vieira de Campos a,*, Marcelo Park b, David Souza Gomez a, Marcus Castro Ferreira a, Luciano Cesar Pontes Azevedo b a

Intensive Care Unit, Burn Unit, Hospital das Clı´nicas, University of Sa˜o Paulo Medical School, Sa˜o Paulo, Brazil Intensive Care Unit, Emergency Department, Hospital das Clı´nicas, University of Sa˜o Paulo Medical School, Sa˜o Paulo, Brazil b

article info

abstract

Article history:

Introduction: To characterize the evolution of clinical and physiological variables in severe

Accepted 30 March 2014

adult burn patients admitted to a Brazilian burn ICU, we hypothesized that characteristics of

Keywords:

Methods: A five-year observational study was carried out. The clinical characteristics,

survivors are different from non-survivors after ICU admission. Intensive care unit

physiological variables, and outcomes were collected during this period.

Burns

Results: A total of 163 patients required ICU support and were analyzed. Median age was 34

Outcomes assessment

[25,47] years. Total burn surface area (TBSA) was 29 [18,43]%, and hospital mortality rate was

Health status indicators

42%. Lethal burn area at which fifty percent of patients died (LA50%) was 36.5%. Median

Fluid balance

SAPS3 was 41 [34,54]. Factors associated with hospital mortality were analyzed in three steps, the first incorporated ICU admission data, the second incorporated first day ICU data, and the third incorporated data from the first week of an ICU stay. We found a significant association between hospital mortality and SAPS3 [OR(95%CI) = 1.114(1.062–1.168)], TBSA [OR(95%CI) = 1.043(1.010–1.076)], suicide attempts [OR(95%CI) = 8.126(2.284–28.907)], and cumulative fluid balance per liter within the first week [OR(95%CI) = 1.090(1.030–1.154)]. Inhalation injury was present in 45% of patients, and it was not significantly associated with hospital mortality. Conclusions: In this study of an ICU in a developing country, the mortality rate of critically ill burn patients was high and the TBSA was an independent risk factor for death. SAPS3 at admission and cumulative fluid balance in the first seven days, were also associated with unfavorable outcomes. The implementation of judicious fluid management after an acute resuscitation phase may help to improve outcomes in this scenario. # 2014 Elsevier Ltd and ISBI. All rights reserved.

* Corresponding author at: Burn Unit, 8th Floor, Hospital das Clı´nicas, Av. Ene´as de Carvalho Aguiar, 255, ZIP 05403-000, Sa˜o Paulo, SP, Brazil. Tel.: +55 11 2661 7221; fax: +55 11 2661 7221. E-mail address: [email protected] (E.V. de Campos). Abbreviations: ICU, intensive care unit; TBSA, total burn surface area; LA50%, lethal burn area in which fifty percent of the patients died; APACHE, acute physiology and chronic health evaluation; SAPS, severity acute physiological score; ABSI, abbreviated burn severity index; SBE, Standard Base Excess; SvO2, venous oxygen saturation; RRT, need for renal replacement therapy; SOFA, sequential organ failure assessment; ROC, receiver operating characteristic; VIF, variance inflation factor. http://dx.doi.org/10.1016/j.burns.2014.03.022 0305-4179/# 2014 Elsevier Ltd and ISBI. All rights reserved.

Please cite this article in press as: de Campos EV, et al. Characterization of critically ill adult burn patients admitted to a Brazilian intensive care unit. Burns (2014), http://dx.doi.org/10.1016/j.burns.2014.03.022

JBUR-4340; No. of Pages 10

2

burns xxx (2014) xxx–xxx

1.

Introduction

Burn is an important cause of morbidity and mortality. Severe cases carry a high risk of physical, emotional, cultural and economic sequelae, not only for the victims but also for society [1]. Burns are common and frequently involve small areas needing only ambulatory treatment. However, more complex patients require hospitalization and intensive care unit (ICU) support [2]. The characteristics of adult patients with burns admitted to the hospital are well described [1,3]. By contrast, data on adult burn patients who need ICU admission are scarce [4–6]. Moreover, very little data on the outcomes of critically ill adult burn patients from low-middle income countries are available. Only one study has explored these features from resource-constrained settings, and this analysis included predominantly very young, non-ventilated patients [6]. Some reasons related to the lack of ICU data on adult burn patients may include the rarity of great thermal injuries [7], burn prevention campaigns [8], the absence of dedicated reporting about burn patients from mixed ICUs, and an improvement in early specific care in burn units that avoids complications and precludes ICU admissions [9]. Knowledge regarding the clinical characteristics, risk factors for mortality, and outcomes of severe burn patients requiring intensive care in low-middle income countries is essential to assist in the decision-making process for care improvement and better resource allocation. Thus, the aim of this study was to characterize severe adult burn patients admitted to a Brazilian burn ICU, with the hypothesis that survivors diverge from non-survivors early in ICU admission. Therefore, the parameters related to hospital mortality were analyzed to improve potential future interventions.

2.

Methods

2.1.

Patients’ enrollment

Data from burn patients who were more than 16 years old and admitted to the burn intensive care unit (a four beds ICU) of a tertiary university hospital in Sa˜o Paulo – Brazil (Hospital das Clı´nicas – University of Sa˜o Paulo) from May 2005 to April 2010 were retrieved from a prospectively observational database. Of 710 patients admitted to the burn unit, 163 (23%) patients were

admitted to the burn ICU (Fig. 1). The criteria used to admit patients into the ICU were: a total burn surface area (TBSA)  20%; inhalation injury; a need for mechanical ventilation; electrical injury; burn shock; associated severe trauma; and severe renal failure. The study was approved by the local Hospital Ethics Committee (Protocol n8 235.362) and written informed consent was waived due to the observational nature of the data collection. Usual care, as described below, was provided to the patients during the study.

2.2.

Study definitions

 Total burn surface area (TBSA): second and third degree burn area extension estimated according to the Lund and Browder diagram [10].  Lethal burn area in which fifty percent of patients died (LA50%): calculated based on the binary logistic regression model, using mortality as the dependent variable and TBSA as the independent variable [11].  Inhalation injury: exposure to fire and smoke in an airtight environment in combination with carbon coating on the face, facial burns, singed nasal vibrissae, singed eyebrows, carbon in the tracheal aspirates or sputum, upper airway edema, a brassy cough and hoarseness, and bronchoscopic findings of airway edema, inflammation, mucosal necrosis, charring and soot [12].  Associated trauma: bone fractures, brain injury and chest and/or abdominal blunt trauma.  Body weight: in patients who were able to speak, the actual weight was self-reported, while in the others the weight was estimated according to the gender and height.  Abbreviated burn severity index (ABSI): This is a specific score that has been developed to predict the hospital mortality rate of burn patients. This consists of epidemiologic variables, such as age and gender, associated with the characteristics of burn trauma, such as inhalation injury, TBSA and the presence of a full thickness burn [13].  Ryan’s score: This is a specific score that has been developed to predict the hospital mortality rate of burn patients. This consists of three clinical variables: age greater than 60 years old, a TBSA greater than 40%, and the presence of inhalation injury [14].  Simplified Acute Physiology Score (SAPS) 3: This is a score that has been developed to predict the hospital mortality rate in patients admitted to the general ICU. It uses epidemiological variables including the origin of the patient

Fig. 1 – Flowchart of the study. Please cite this article in press as: de Campos EV, et al. Characterization of critically ill adult burn patients admitted to a Brazilian intensive care unit. Burns (2014), http://dx.doi.org/10.1016/j.burns.2014.03.022

JBUR-4340; No. of Pages 10 burns xxx (2014) xxx–xxx

before an admission to the ICU, physiological variables and laboratory data from the first hour in the ICU [15].  Acute physiology and chronic health evaluation (APACHE) II: This is a score that has been developed to predict the hospital mortality rate of patients admitted to the general ICU. It uses epidemiological variables, physiological variables and laboratory data from the first twenty four hours after the ICU admission [16].

2.3.

Burn patients’ treatment protocol

In addition to the standard ICU support, the patients also received specific general burn care, which included individualized rooms, early fluid resuscitation according to the Modified Brooke formula adding fluid challenges in the presence of hyperlactatemia (lactate > 2 mmol/L) and oliguria (urinary output < 0.5 ml/kg/h), surgical excision and grafting of the burn wound (limited by local resources, in which there are no available synthetic skin substitutes and skin bank supplies are insufficient to demand), daily dressing changes, topical chemotherapy (silver sulfadiazine), vigorous enteral nutrition based on the Curreri approach (estimated through the TBSA and body weight) [17], individualized analgesia and the early initiation of rehabilitation.

2.4.

Data collected

The data recorded at admission were age, gender, SAPS 3 and APACHE II scores, and the presence of hypotension (systolic systemic arterial pressure < 120 mmHg). The burn characteristics collected at admission included the ABSI, Ryan’s score, TBSA, mechanism of thermal injury (fire burn, scald, electrical, chemical or flash burn), presence of inhalation injury, interval from the time of injury to the ICU admission, presence of associated trauma, injury resulting from a suicide attempt and escharotomy promptly needed at admission. The comorbidities were also registered. During the ICU evaluation and treatment, data were collected until death, discharge, the fourth day [lactate, Standard Base Excess (SBE), and SvO2] or the seventh day after admission (all other variables). The support variables that were monitored and recorded daily were the dobutamine and/or norepinephrine use, cumulative fluid balance, diuresis, need for mechanical ventilation, need for renal replacement therapy (RRT), creatinine level, lactate level, SBE, venous hemoglobin oxygen saturation (collected from the femoral vein after a venous catheter insertion), platelet count, P/F ratio and total Sequential Organ Failure Assessment (SOFA) score with its components [18]. The clinical outcomes collected were the ICU length-of-stay, hospital length-of-stay, and ICU and hospital mortality.

2.5.

Statistical analysis

The data distribution was analyzed using the Kolmogorov– Smirnov goodness-of-fit model. The data were predominantly non-normally distributed, therefore the quantitative variables are shown as the median [P25th, P75th] and were analyzed using a Mann–Whitney test for two non-paired variables and a Friedman’s test for more than two paired

3

variables. The post hoc analyses were performed using Dunn’s test. To avoid a type II error when comparing the survivors vs non-survivors for the first four days, the P value used for each daily comparison was 0.0125 (Bonferroni’s correction for multiplicity). For the other statistical tests we used a P value < 0.05 to indicate significance. A Chi-square test or Fischer’s exact test was used to compare categorical variables as appropriate. The accuracy of the data associated with hospital mortality was measured using the area under the receiver operating characteristic (ROC) curve. The best value of a given variable, and the respective sensitivity and specificity, were retrieved using the Youden’s index. The data for all the patients were recorded and categorized according to survival; the variables with P < 0.05 in the single survivors vs non-survivors comparisons (univariate analysis) were used in the multivariate model. The details of the multivariate analysis are shown in the supplementary material. Briefly, the single and multiple collinearity of the data used as independent variables were measured using Spearman’s correlation (collinearity was taken to be at an R > 0.85) and the variance inflation factor (VIF, collinearity was taken to be at a VIF > 2.5), respectively. The variables with high collinearity were excluded from the analysis according to their clinical relevance. The multivariate analysis was performed using a hierarchical binary logistic regression model. This methodology is explained in detail in the supplementary material. The backward elimination was performed twice. Firstly, it was performed manually, with the variable extractions based on the significance of the association and the odds ratio (each step in the model was performed three times with a bootstrapping of 2500 replacements in each run); secondly, it was performed automatically based on the likelihood ratio variation. The first hierarchy included the ICU admission variables; the second hierarchy included the laboratory and clinical data at the first day of ICU; and the third hierarchy analyzed the first week of an ICU stay, using the worst value of the organ failure quantification, the cumulative fluid balance and the need for ICU support. The next hierarchy preserved the previous one, inserting the previous hierarchy values as unchanging variables into the current analysis (as enter mode). The automated backward elimination used a stepwise probability of P < 0.05 as entry criteria and a P < 0.1 as exclusion criteria. The analysis was carried out using the Statistical Package for Social Sciences 19.0 (SPSS1 Inc., Chicago, IL, USA), and the data were graphed using the SigmaPlot1 9.0 software (Systat software, San Jose, CA, USA).

3.

Results

All the ICU-admitted patients were included in the analysis. The characteristics of the patients are shown in Table 1. Notably, the patients were young and the non-survivors were older than the survivors. The survivors were more frequently men with lower SAPS 3 scores and a lower TBSA. An inhalation injury was more common in the non-survivors. The LA50% was 36.5% and the greatest TBSA of all the patients who survived in our study was 57%.

Please cite this article in press as: de Campos EV, et al. Characterization of critically ill adult burn patients admitted to a Brazilian intensive care unit. Burns (2014), http://dx.doi.org/10.1016/j.burns.2014.03.022

JBUR-4340; No. of Pages 10

4

burns xxx (2014) xxx–xxx

Table 1 – General characteristics of the patients according to survival. All patients (n = 163) Admission data Age – yo Male gender – n(%) SAPS 3 – points APACHE II – points ABSI – points Ryan score – points Total SOFA at admission – points Systolic pressure < 120 mmHg – n(%)a Burn/trauma characteristics Total burn surface area – (%) 20 – n(%) 20–40 – n(%) 40 – n(%) Fire burn – n(%) Scald – n(%) Electrical – n(%) Chemical – n(%) Flash burn – n(%) Inhalation injury – n(%) Interval to admission – h Associated trauma – n(%) Suicide attempt – n(%) Escharotomy – n(%)b Comorbidities Alcohol abuse or illicit drugs – n(%) Arterial hypertension – n(%) Diabetes – n(%) Stroke – n(%)

34 116 41 11 8 1 3 64

[25,47] (71) [34,54] [6,16] [6,9] [0,2] [1,7] (39)

29 [18,43] 59 (36) 59 (36) 45 (28) 121 (74) 7 (4) 28 (17) 3 (2) 4 (3) 74 (45) 3.0 [1.0,14.0] 12 (7) 32 (20) 53 (33) 27 7 3 3

(17) (4) (2) (2)

Survivors (n = 94) 33 73 37 7 6 0 1 24

Non-survivors (n = 69)

[25,44] (78) [32,43] [5,11] [5,7] [0,1] [1,4] (38)

37 43 54 16 9 2 7 40

P value#

[26,53] (62) [45,66] [11,21] [8,12] [1,2] [4,9] (62)

0.098 0.025

Characterization of critically ill adult burn patients admitted to a Brazilian intensive care unit.

To characterize the evolution of clinical and physiological variables in severe adult burn patients admitted to a Brazilian burn ICU, we hypothesized ...
944KB Sizes 0 Downloads 4 Views