The Advanced Trauma Life Support system classifies the severity of shock. The aim of this study is to test the validity of this classification.Admission physiology, injury and outcome variables from adult injured patients presenting to hospitals in England and Wales between 1989 and 2007 and stored on the Trauma Audit and Research Network (TARN) database, were studied. Patients were divided into groups representing the four ATLS classes of shock, based on heart rate (HR) systolic blood pressure (SBP), respiratory rate (RR) and Glasgow Coma Score (GCS). The relationships between variables were examined by classifying the cohort by each recorded variable in turn and deriving the median and interquartile range (IQR) of the remaining three variables. Patients with penetrating trauma and major injuries were examined in sub-group analyses.In blunt trauma patients grouped by HR, the median SBP decreased from 128mmHg in patients with HR<100BPM to 114mmHg in those with HR>140BPM. The median RR increased from 18 to 22bpm and the GCS reduced from 15 to 14. The median HR in hypotensive patients was 88BPM compared to 83BPM in normotensive patients and the RR was the same. When grouped by RR, the HR increased with increasing RR but there were no changes in SBP.In trauma patients there is an inter-relationship between derangements of HR, SBP, RR and GCS but not to the same degree as that suggested by the ATLS classification of shock.