Models of Certain Types of Surgical Shock Based on the Assessment of Selected Critical Parameters
The aim of the study was to devise models that describe three types of surgical shock based on a set of simple biochemical and clinical parameters and establish a method of assigning new patients to each surgical shock scenario.
Material and methods. Prospective analysis included patients hospitalized in IInd Chair of Surgery from 2001 to 2005 who belonging to the following groups: multiple injuries (at least 7 points according to LSO), upper GI bleeding (in patients requiring blood transfusions during first day) and severe acute pancreatitis (at least 3rd degree according to Trapnell's scale). A set of nineteen parameters was evaluated on the first, fourth and seventh day in every patient. Discrimination analysis was used for statistical analysis with calculation of Mahalanobis squared distances between groups that reflect their reciprocal differences. Discrimination functions were calculated allowing the assignment of a new observation to one of the models. Eventually, reliability of categorizing patients to the studied groups was evaluated.
Results. Differences among the groups - reflected by Mahalanobis squared distances - proved statistically significant in every studied day. Overall, the ratio of proper classifications according to discrimination functions equaled about 87%. The most common mistake in categorizing was observed in groups of multiple injuries and acute pancreatitis - due to similar pathophysiological processes leading through SIRS to MOF.
Conclusions. The abovementioned results indicate that the presented models can be successfully used in diagnostic processes, especially in emergency cases. The set of clinical and biochemical parameters used is simple and easy to obtain even in non-specialized centers.