Current techniques for assessment of chest trauma rely on clinical diagnoses or scoring systems. However, there is no generally accepted standard for early judgement of the severity of these injuries, especially in regards to related complications. This drawback may have a significant impact on the management of skeletal injuries, which are frequently associated with chest trauma. However, no convincing conclusions can be determined until standardization of the degrees of chest trauma is achieved. We investigated the role of early clinical and radiologic assessment techniques on outcome in patients with blunt multiple trauma and thoracic injuries and developed a new scoring system for early evaluation of chest trauma. A retrospective investigation was performed on the basis of 4,571 blunt polytrauma (Injury Severity Score [ISS] > or = 18) patients admitted to our unit. Inclusion criteria were treatment of thoracic injury that required intensive care therapy, initial Glasgow Coma Scale score greater than 8 points, and no local or systemic infection. Patients with thoracic trauma and multiple associated injuries (ISS > or = 18) were included. In all patients, the association between various parameters of the thoracic injuries and subsequent mortality and morbidity was investigated. A total of 1,495 patients fulfilled the inclusion criteria. Patients' medical records and chest radiographs were reevaluated between May 1, 1998, and June 1, 1999. The association between rib fractures and chest-related death was low (> three ribs unilateral, mortality 17.3%, odds ratio 1.01) unless bilateral involvement was present (> three ribs bilateral, mortality 40.9%, odds ratio 3.43). Injuries to the lung parenchyma, as determined by plain radiography, were associated with chest-related death, especially if the injuries were bilateral or associated with hemopneumothorax (lung contusion unilateral, mortality 25.2%, odds ratio 1.82; lung contusion bilateral + hemopneumothorax, mortality 53.3%, odds ratio 5.1). When plain anteroposterior chest radiographs were used, the diagnostic rate of rib fractures (< or = three ribs) increased slightly, from 77.1% to 97.3% during the first 24 hours of admission. In contrast, pulmonary contusions were often not diagnosed until 24 hours after admission (47.3% at admission, 92.4% at 24 h, p = 0.002). A new composite scoring system (thoracic trauma severity score) was developed that combines several variables: injuries to the chest wall, intrathoracic lesions, injuries involving the pleura, admission PaO2/FIO2 ratio, and patient age. The receiver operating characteristic curve demonstrated an adequate discrimination, as demonstrated by a value of 0.924 for the development set and 0.916 for the validation set. The score was also superior to the ISS (0.881) or the thorax Abbreviated Injury Score (0.693). Radiographically determined injuries to the lung parenchyma have a closer association with adverse outcome than chest-wall injuries but are often not diagnosed until 24 hours after injury. Therefore, clinical decision making, such as about the choice of surgery for long bone fractures, may be flawed if this information is used alone. A new thoracic trauma severity score may serve as an additional tool to improve the accuracy of the prediction of thoracic trauma-related complications.