Penetrating chest injury: who warrants aggressive treatment?

To determine the degree of severity in penetrating chest injuries that predicts survival, we conducted a logistic regression analysis. All patients suffering penetrating chest injuries (n = 310) admitted to an urban level I teaching hospital in the USA between January 1993 and December 1994 were evaluated. The Injury Severity Score (ISS), Glasgow Coma Scale (GCS), Trauma Score (TS), and Revised Trauma Score were used to compare injury survivors with nonsurvivors. We used the trauma scores to create a logit to predict the outcome among 160 patients in 1993 and tested the validity of this logit in another 150 patients in 1994. With death = 0, survival = 1, the equation lnPd/Ps = b0 + b1 ISS + b2 GCS + b3 TS was obtained from logistic regression, where b0 was the constant of the equation and b1, b2, and b3 were the coefficients of ISS, GCS, and TS, respectively. A logit score greater than 0.5 was found to be predictive of death with a sensitivity of 80.0%, a specificity of 97.5%, and an accuracy of 94.0%. Aggressive resuscitation should be aimed at patients with a logit score greater than 0.5 to reduce mortality. This knowledge may aid in the management of patients with severe chest injuries.