Revision of the multiple organ failure score

The multiple organ failure (MOF) score published by Goris et al. in 1985 was one of the first attempts to quantify severity of organ dysfunction and failure based on expert opinion in surgical intensive care unit patients. Fifteen years later a reassessment of this score is mandatory.

Daily MOF scores were documented in patients admitted to the surgical ICUs in Nijmegen (NL) and Cologne (D). Patients with an ICU stay < or = 3 days were excluded. Organ dysfunction (1 point) and organ failure (2 points) were recorded for the following organ systems: lung, heart, kidney, liver, blood, gastrointestinal tract (GI), and central nervous system (CNS). Maximum scores were computed, and logistic regression analysis was used to optimize point weights for each organ system. Predictive power was analyzed using receiver operating characteristic (ROC) curves.

In all, 147 patients, mean age 56 years, were included with a total of 2,354 observation days. Hospital mortality was 30.6%. GI failure was present on only 3.3% of days, without impact on mortality. Valid evaluation of CNS was impossible in most cases due to sedation and ventilation. Reweighting of the score items revealed only marginal improvements in prediction. Mortality consistently increased with increase in number of failed organs. This phenomenon was even more pronounced in older patients, e.g., 55% mortality (age > or = 60) versus 0% (age < 60) with two failing organs.

Due to problems in definition and assessment (reliability) CNS and GI should not be considered in future assessments of the MOF score. The original point weights in the remaining five organ systems provide a valid and reliable risk stratification, at least in surgical ICU patients.