Pre-hospital assessment and management of major trauma is challenging, and this research highlights some of the complexities confronting ambulance staff when encountering these events. The purpose of this retrospective study was to validate the pre-hospital trauma life support (PHTLS) classification of hypovolaemic shock, derived from the advanced trauma life support (ATLS) guidelines.
Data on 46 689 adult patients (2002 until 2011) were extracted from a German trauma register and were classified into reference ranges reflecting the PHTLS classification of hypovolaemic shock: heart rate (HR), systolic blood pressure (SBP) and Glasgow coma scale (GCS).
The first stage of analysis involved grouping patients by a combination of all three of these factors as suggested by PHTLS; in the next stage patients were classified by one parameter (HR, SBP or GCS) only and alterations in the remaining two parameters were assessed.
Overall, the results show substantial deficits in the PHTLS classification of hypovolaemic shock when risk-stratifying trauma patients.
In summary, the vast majority of patients (97.9%) fell into class I or II shock. Only 12 432 (26.5%) patients could be correctly classified according to PHTLS if all three criteria were taken into account. Accuracy decreased to 12.2% in patients in traumatic brain injury (TBI). There was no statistically remarkable difference in accuracy between penetrating and blunt trauma (28.3% versus 26.6%).
It was observed that the PHTLS classification tended to over-emphasise the significance of HR in shock. The association between tachycardia and hypotension was found to be weak in all groups. When patients were classified by HR, mean SBP dropped slightly from 125 (±33) mmHg in class I shock, to 111 (±43) mmHg in class IV decompensated shock—above the PHTLS guidelines for that category of patient.
An interesting discussion emerged around the findings relating to GCS where the authors suggest that PHTLS may underestimate ‘mental disability’ linked to hypovolaemic shock. If grouped by GCS there appeared to be an association between a decrease in GCS and a decline in SBP. Similarly, when classified by SBP on scene, normotensive patients (SBP≥110 mmHg) had a median GCS of 14 (11–15), whereas in hypotensive patients (SBP<90 mmHg), a median GCS of 8 (3–14) was observed.
The authors speculate only briefly on the reason behind the inaccuracies, including an underestimation of the compensatory mechanisms in severe injury and of the physiological effects of pain, anxiety and drugs on vital signs.
A growing body of evidence is revealing inadequacies in the traditional non-invasive methods of monitoring and classifying trauma patients. Other studies have found higher fidelity in gauging hypovolaemia through the measurement of pre-hospital lactate levels.
The researchers clearly articulate some limitations of this study including those associated with retrospective research. In addition, they recognise that the trauma database that they used does not include information about the patients' previous use of medications, e.g. antihypertensive drugs or consumption of alcohol etc., all of which could impact on the parameters assessed; neither does it include all components of the PHTLS classification system, e.g. urine output and pulse pressure.
However, notwithstanding these limitations, this is an important step in the direction of evidence-based pre-hospital trauma care. It comes at a time when long-held beliefs are being debunked: for instance the infamous ‘golden-hour’, now shown through several studies to have little grounding in evidence.
The implications of this and similar research on the management and treatment of pre-hospital trauma are profound. It shows that a high index of suspicion is needed when interpreting vital signs of patients involved in major trauma. The authors conclude by identifying a need for further validation using other trauma databases recommending that, ultimately, a critical appraisal of the PHTLS classification system may be necessary.