Comparison of multiorgan dysfunction (MOD) scores in prediction of 1 -year mortality of ICU patients

Introduction

MOD and secondary infections are main causes of death in intensive care unit. The developed scores and models for prediction of mortality, as APACHE II, III and MPM II predict outcome quite accurately in a large population but are rather inaccurate and should be applied with caution for a single ICU patient. Although the basic idea in development of MOD scores like MODS, LOD and SOFA was not to predict mortality, but to describe the number, severity and progression of organ failures, these scores seem to correlate with mortality. The objective of this study was to compare these new scores in prediction of 1-year mortality of ICU patients.

Methods

In year 1995 our mixed 10-bed ICU had 592 admissions. 333 patients were randomly chosen for this study. Data of all admissions was collected partly prospectively, but the MOD scores were calculated retrospectively because of availability only since 1996. All deaths were verified from the Finnish National Registry at June 1997. 1-year mortality rate for various values of APACHE III, MODS, LOD and SOFA scores were evaluated and compared by calculating area under receiver operating curves (AUCs).

Results

Areas under ROC (AUC) were 0.7817 for APACHE III, 0.7570 for LOD (day 1), 0.7226 for MODS (day 1) and 0.7215 for SOFA (day 1) in prediction of 1-year mortality.

Discussion

There is no clear consensus of method to be used in evaluation of MOD. Several scoring systems, as MODS, LOD and SOFA, have been presented, but they have not been properly compared in different patient populations. All of the scores clearly correlated with 1-year mortality in this study. The LOD score (at ICU day 1) had the highest predictive power with the AUC nearly equal to that of APACHE III.

Conclusions

In addition to usefulness of MODS, LOD and SOFA scores in assessment of multiorgan dysfunction, these scores may be used in prediction of mortality. Further studies are needed to evaluate the differences of these scores in this respect, as well as the possible advantage of combining a multiorgan dysfunction score at different timepoints (for example ICU days 3, 5 and 7), or its change, with baseline APACHE III score in prediction of mortality.