The Acute Physiology and Chronic Health Evaluation II (APACHE II) score has never been validated to risk-adjust between critically ill trauma (TICU) and general surgical (SICU) intensive care unit patients, yet it is commonly used for such a purpose. To study this, we evaluated risk of death in TICU and SICU patients with pneumonia. We hypothesized that mortality for a given APACHE II would be significantly different and that using APACHE II to directly compare TICU and SICU patients would not be appropriate. We conducted a retrospective review of patients admitted to the TICU or SICU at a tertiary medical center over an 18-month period with pneumonia. Admission APACHE II scores, in-hospital mortality, demographics, and illness characteristics were recorded. One hundred eighty patients met inclusion criteria, 116 in the TICU and 64 in the SICU. Average APACHE II scores were not significantly different in the TICU versus SICU (25 vs 24; P = 0.4607), indicating similar disease severity; overall mortality rates, however, were significantly different (24 vs 50%; P = 0.0004). Components of APACHE II, which contributed to this mortality differential, were Glasgow Coma Score, age, presence of chronic health problems, and operative intervention. APACHE II fails to provide a valid metric to directly compare the severity of disease between TICU and SICU patients with pneumonia. These groups represent distinct populations and should be separated when benchmarking outcomes or creating performance metrics in ICU patients. Improved severity scoring systems are needed to conduct clinically relevant and methodologically valid comparisons between these unique groups.