Comparison of the local and systemic effects of sepsis in predicting survival

Progress in the study of sepsis has been hampered by the lack of a suitable system for grading its severity. Systems suggested for scoring sepsis have been based either on its systemic effects (APACHE II) or on a mixture of local and systemic variables (sepsis score). The APACHE II and sepsis scores were applied to patients with intra-abdominal sepsis of more than 3 days' duration, to determine if local or systemic factors were more important in predicting survival. Of 45 patients studied, 14 died. The sepsis score for non-survivors (median 21.5, range 11-32) was significantly higher than for survivors (median 14, range 10-26, P less than 0.05). There was overlap between the two groups, such that an individual score had no predictive value. The local component of the sepsis score was not significantly increased in non-survivors (P less than 0.05), but the systemic component was (P less than 0.05). The APACHE II score for non-survivors (median 24, range 15-38) was significantly higher than for survivors (median 12, range 3-21), and correctly identified 13 of the 14 fatalities. Both the systemic and non-systemic components (age and chronic disease) were significantly higher among the latter. The APACHE II was more effective than the sepsis score in predicting survival. We conclude that any system used for scoring sepsis should be based on systemic rather than local effects. At present the APACHE II score is preferred.