- Title
- Post-injury multiple organ failure: epidemiology, prediction modelling, and score comparison in an Australian setting
- Creator
- Dewar, David Craig
- Relation
- University of Newcastle Research Higher Degree Thesis
- Resource Type
- thesis
- Date
- 2016
- Description
- Research Doctorate - Doctor of Philosophy (PhD)
- Description
- Post-‐injury multiple organ failure (MOF) is a morbid, lethal syndrome caused by systemic inflammation following significant trauma and haemorrhagic shock. Due to developments in trauma surgery and critical care medicine patients are able to survive previously mortal injuries, thereby allowing the syndrome of MOF to develop. MOF continues to represent the majority of late trauma deaths, and provides a major cost to advanced health care systems. Primary Aims: The primary aim of this thesis was to define the epidemiology of MOF in Australia. There has been no investigation of MOF within Australia previous to this work, representing a significant gap in the literature. Australia has a unique mature health care system, with specific injury patterns, high levels of road trauma and low levels of penetrating trauma, and significant geographical distances that patients may have to travel to seek care. It was unclear if previously published international epidemiological data would be comparable in an Australian context. The John Hunter Hospital MOF database was created to identify and describe cases of MOF arising at a Level 1 trauma centre. The John Hunter Hospital has the highest volume of trauma admission within the state of New South Wales (the most populous state of Australia), and 420 patients with an injury severity score (ISS) greater than 15 are treated at John Hunter Hospital annually. Patients were included based on survival for 48 h, injury severity, and intensive care unit (ICU) admission. Patients were excluded if they had a significant head injury (abbreviated injury scale [AIS]>2), significant burns or hanging injury. MOF was defined using the Denver MOF score to allow for comparison to the longest running MOF database. Assessments were made at 12 months and five years. The incidence of MOF in this high risk cohort was 15 per cent and MOF related mortality was 24 per cent. These results are world leading for both incidence and MOF related mortality, and provide evidence that severely injured patients have good outcomes within our trauma system. Further, this allows comparison of our cohort to other large international data sets. Secondary Aims: This thesis had two secondary aims. The first secondary aim was to develop a clinical tool that could be used to predict MOF as early in the clinical course as possible. The second secondary aim was to comparatively assess the utility of two MOF scores to predict objective outcomes. Multiple prediction models have been developed to identify which patients are more likely to develop MOF. Potential prediction markers were identified in the development of the MOF data set and were based on previous investigations and expert opinion.Univariate and multivariate analyses were performed on these prediction markers using the five-year cohort to develop models that could be used to predict MOF. Two models were developed, at two distinct time periods. The first model could be applied in the emergency department and used two variables (age >65 years old and platelet count on admission of < 150/10⁹L); the area under the curve (AUC) for this model was 0.71. One point was scored for each variable. Patients who scored 2 points had a probability of MOF of 40 per cent, patients who scored 1 had a probability of 26 per cent, and patients who scored 0 had a probability of 8 per cent. Multivariate analysis also was used to develop a model to be used at 24 hours post-injury. Final variables in the model were (1 point for age >65 years old and/or platelet count on admission of < 150/10⁹L, 4 points for maximum creatinine within 24 hours of > 150/10⁹L, and 5 points for minimum bilirubin > 10/10⁹L), the AUC for this model was 0.84. Patients who scored <2 points had a probability of MOF of 0 per cent, patients who scored 2–5 points had a probability of 10 per cent and patients who scored > 5 points had a probability of 38 per cent. None of the previously defined prediction markers were significant in our cohort. Multiple composite scoring systems have been used to define MOF. This creates significant difficulties within the literature and it is extremely difficult to compare published results. There is a very large range of reported incidence of MOF due to the abundance of definitions and disparate cohort inclusion and exclusion criteria. Two of the most widely used scoring systems are the Denver MOF score and the sequential organ failure assessment (SOFA). Both scores have been validated in trauma patients; however, previous to our investigation there has been no head-to-head comparison of the two scoring systems. Both the Denver and SOFA scores were collected prospectively on all patients within the cohort, and the five-year data set was analysed to determine the performance of both scores with respect to mortality, ICU length of stay (ICU LOS) and ventilator days. Sensitivity, specificity and receiver operating characteristic (ROC) curves were created for both the SOFA and Denver scores, and cut points to define MOF were analysed for performance. There were significant differences between the performances of the two scores. The SOFA score had strong sensitivity and poor specificity, while the Denver score had poor sensitivity and strong specificity to predict mortality. There was no difference in the AUC for mortality. This finding was replicated when both scores were compared against ICU LOS and ventilator days: the SOFA score had high sensitivity and low specificity, while the Denver score had low sensitivity and high specificity, however the AUC for the SOFA score was higher when assessed against ICU LOS and ventilator days. Despite the performance improvement the SOFA score was challenging to apply clinically. The SOFA score requires an assessment of the Glasgow Coma scale (GCS), which is not always available in the ICU setting. Difficulties arose when patients were intubated and ventilated for their injuries, or experienced respiratory failure secondary to MOF and so an accurate central nervous system score was not always available. This created significant scoring inaccuracy as patients with MOF may have remained ventilated for multiple consecutive days.
- Subject
- trauma; multiple organ failure; thesis by publication
- Identifier
- http://hdl.handle.net/1959.13/1321976
- Identifier
- uon:24484
- Rights
- Copyright 2016 David Craig Dewar
- Language
- eng
- Full Text
- Hits: 878
- Visitors: 1169
- Downloads: 323
| Thumbnail | File | Description | Size | Format | |||
|---|---|---|---|---|---|---|---|
| View Details Download | ATTACHMENT01 | Thesis | 6 MB | Adobe Acrobat PDF | View Details Download | ||
| View Details Download | ATTACHMENT02 | Abstract | 145 KB | Adobe Acrobat PDF | View Details Download |