To develop an objective scale to measure the severity of the multiple organ dysfunction syndrome as an outcome in critical illness.
Systematic literature review; prospective cohort study.
Surgical intensive care unit (ICU) of a tertiary-level teaching hospital.
All patients (n equals 692) admitted for more than 24 hrs between May 1988 and March 1990.
Computerized database review of MEDLINE identified clinical studies of multiple organ failure that were published between 1969 and 1993. Variables from these studies were evaluated for construct and content validity to identify optimal descriptors of organ dysfunction. Clinical and laboratory data were collected daily to evaluate the performance of these variables individually and in aggregate as an organ dysfunction score. Seven systems defined the multiple organ dysfunction syndrome in more than half of the 30 published reports reviewed. Descriptors meeting criteria for construct and content validity could be identified for five of these seven systems: a) the respiratory system (PO2 /FIO2 ratio); b) the renal system (serum creatinine concentration); c) the hepatic system (serum bilirubin concentration); d) the hematologic system (platelet count); and e) the central nervous system (Glasgow Coma Scale). In the absence of an adequate descriptor of cardiovascular dysfunction, we developed a new variable, the pressure-adjusted heart rate, which is calculated as the product of the heart rate and the ratio of central venous pressure to mean arterial pressure.
These candidate descriptors of organ dysfunction were then evaluated for criterion validity (ICU mortality rate) using the clinical database.From the first half of the data-base (the development set), intervals for the most abnormal value of each variable were constructed on a scale from 0 to 4 so that a value of 0 represented essentially normal function and was associated with an ICU mortality rate of less than 5%, whereas a value of 4 represented marked functional derangement and an ICU mortality rate of more than equals 50%. These intervals were then tested on the second half of the data set (the validation set). Maximal scores for each variable were summed to yield a Multiple Organ Dysfunction Score (maximum of 24). This score correlated in a graded fashion with the ICU mortality rate, both when applied on the first day of ICU admission as a prognostic indicator and when calculated over the ICU stay as an outcome measure. For the latter, ICU mortality was approximate 25% at 9 to 12 points, 50% at 13 to 16 points, 75% at 17 to 20 points, and 100% at levels of more than 20 points. The score showed excellent discrimination, as reflected in areas under the receiver operating characteristic curve of 0.936 in the development set and 0.928 in the validation set. The incremental increase in scores over the course of the ICU stay (calculated as the difference between maximal scores and those scores obtained on the first day [i.e., the Delta Multiple Organ Dysfunction Score]) also demonstrated a strong correlation with the ICU mortality rate. In a logistic regression model, this incremental increase in scores accounted for more of the explanatory power than admission severity indices.
This multiple organ dysfunction score, constructed using simple physiologic measures of dysfunction in six organ systems, mirrors organ dysfunction as the intensivist sees it and correlates strongly with the ultimate risk of ICU mortality and hospital mortality. The variable, Delta Multiple Organ Dysfunction Score, reflects organ dysfunction developing during the ICU stay, which therefore is potentially amenable to therapeutic manipulation. As an outcome measure complementary to predictive scores now widely available, such a score may find use in epidemiologic studies of the multiple organ dysfunction syndrome. Moreover, an instrument that can provide an objective measure of the severity of organ dysfunction at the time of ICU admission and that can quantify subsequent deterioration over the course of the ICU stay may prove useful as an alternative end point for clinical trials involving critically ill patients.
(Crit Care Med 1995; 23:1638-1652)
From the Department of Surgery, University of Toronto, Toronto, ON, Canada (Dr. Marshall); Department of Medicine, McMaster University, Hamilton, ON, Canada (Dr. Cook); Department of Surgery, McGill University, Montreal, PQ, Canada (Dr. Christou); Department of Medicine, Vanderbilt University, Nashville, TN (Dr. Bernard); Department of Anesthesia, Hadassah Hebrew University, Jerusalem, Israel (Dr. Sprung); and the Department of Medicine, University of Western Ontario, London, ON, Canada (Dr. Sibbald).
This study was supported, in part, by the Victoria General Hospital, Halifax, NS, Canada, and the Surgical Intensive Care Unit, Toronto Hospital, General Division, Toronto, ON, Canada.
This study was presented, in part, at the 21st Annual Educational and Scientific Symposium of the Society of Critical Care Medicine, San Antonio, TX, May 1992.
Address requests for reprints to: John C. Marshall, MD, Eaton North 9-234, Toronto Hospital, General Division, 200 Elizabeth Street, Toronto, ON M5G 2C4, Canada.