LAUSR

Initial correspondence from Drs. Haas, Karakus, Fijen and de Lange Dear Editor, With great interest, we have read the article by Vincent et al. describing changes in time in the characteristics and outcomes of patients admitted to the intensive care unit in 2002 versus 2012 [1]. The authors show that, in 2012, patients were somewhat older, presented more with septic shock and needed more renal replacement therapy, but mortality declined. These are very important findings, but some limitations to this comparison need to be addressed. First of all, the definition of sepsis differed between the two studies. In the SOAP study, sepsis was defined as having two SIRS criteria plus an infection, whereas in the ICON study, sepsis was defined as infection and at least one organ failure. Although in reality these definitions will substantially overlap, differences caused by selection of the “septic population” cannot be excluded. Indeed, previous studies have shown that applying sepsis two versus sepsis three criteria in the same population led to slightly different selections of patients and different mortality trends in time [2]. Second, the proportion of patients from the different European countries appear to differ between 2002 and 2012, while there were remarkable differences in mortality rates between those countries. The differences between the SOAP and ICON could, at least partially, be attributed to differences in case mix and outcome of included hospitals. Possibly the best way to discern trends in time can be derived from national registries that include all ICU patients [3], although these cannot be extrapolated to other countries.