Objective: The multiple organ dysfunction score (MODS) describes and quanti
fies organ-specific physiology. The objective of this study was to examine
the relation between six components of MODS (cardiovascular, respiratory, r
enal, central nervous system, hepatic, and hematologic) measured at admissi
on to the intensive core unit (ICU) and during the ICU stay, with time to d
eath in the ICU.
Design: Prospective observational cohort study.
Setting. Sixteen Canadian ICUs.
Patients., A total of 1,200 patients were mechanically ventilated for > 48
hrs.
Measurements and Main Results. The six organ systems comprising MODS were m
easured at ICU admission (baseline scores) and daily thereafter. The change
in organ dysfunction each day (serial scores) were calculated as daily com
ponent scores minus the corresponding baseline component scores. In Cox reg
ression analyses, the independent explanatory variables were the MODS compo
nents measured at baseline and serially, and the dependent variable was the
time from admission to ICU mortality. When each organ system was analyzed
individually, both the baseline and serial MODS for the cardiovascular, res
piratory, renal, central nervous system, and hematologic components were si
gnificantly associated with ICU mortality. After adjusting for the serial h
epatic score, the baseline hepatic score was unrelated to mortality. After
adjusting for all baseline and serial MODS components in aggregate, four or
gan systems were significantly associated with ICU mortality: cardiovascula
r (baseline relative risk [RR], 1.5; serial RR, 1.4); respiratory (baseline
RR, 1.4; serial RR, 1.4); renal (baseline RR, 1.3; serial RR, 1.5); and ce
ntral nervous system (baseline RR, 1.6; serial RR, 1.7). We found that the
relative risk of mortality related to organ dysfunction varied significantl
y over time and among organ systems. Baseline respiratory function was not
associated with mortality until the second ICU week (week 1: RR, 1.1 [0.9-1
.4]; week 2 onward: RR, 1.9 [1.5-2.4]); the same was true for the change in
respiratory function as measured by the serial respiratory score (week 1:
RR, 1.2 [1.0-1.5]; week 2 onward: RR, 1.7 [1.4-2.1]). The serial hepatic sc
ore was not associated with mortality until the fourth ICU week (weeks 1-3:
RR, 0,9 [0.7-1.1]; week 4 onward: RR, 1.4 [1.0-2.0]).
Conclusions: Organ dysfunction scores describe physiology at ICU admission
and during ICU stay. Although patterns vary by system, daily MODS component
scores provide additional prognostic value over baseline MODS.