High Mortality in Sepsis Survivors

ABSTRACT & COMMENTARY

Sepsis is usually considered an acute illness, the severity of which correlates with acute mortality. It is apparent that the mortality of the disease is related to the systemic manifestations experienced by the patient rather than the presence of a particular (or any) pathogenic organism. This has led to renaming "sepsis" as the "systemic inflammatory response syndrome" (SIRS) and grading its severity by the degree of physiological decompensation. Most studies of SIRS have focused on the short-term outcome, limiting mortality considerations to 30 days. Quartin et al attempted to identify the long-term effect on survival in patients who initially survived an episode of SIRS.

This study was a continuation of the Veterans Affairs Medical Centers Cooperative Study of Corticosteriods in Systemic Sepsis study carried out between 1983 and 1986. In the study, 1505 patients were identified when they were admitted to a participating hospital having survived an episode of SIRS during the previous year. The control group consisted of the 91,830 patients admitted during the same time period without an episode of SIRS in the previous year. The SIRS episode was classified as uncomplicated sepsis (40%), severe sepsis (45%), and septic shock (15%) according to standard definitions. Demographics and co-morbid conditions were identified from medical records for both groups. Death was the primary outcome measure, and patients were followed for up to eight years.

Using the mortality of the control group patients, a predictive model for death was developed. Applying the model to the SIRS survivors, a 24-28% one-year mortality was expected based on the recorded co-morbidities and demographics. The actual one-year mortality was 61%. SIRS survivors continued to experience higher than expected mortality throughout the eight years of follow up. The relative risk of death ranged from 1.2 to 3.1 times the control group for the remaining eight years of the study. During the first year after SIRS, the relative risk ranged from 16.8 to 3.5 times the control group and was greater with worsening sepsis. Survival from SIRS seems to carry increased mortality risk throughout the patient’s lifetime. (Quartin AA, et al. JAMA 1997;277:1058-1063).

COMMENT BY CHARLES G. DURBIN, Jr., MD, FCCM

This is a very provocative study. SIRS seems to carry an increased mortality risk above other co-morbid factors throughout the survivor’s lifetime. These patients did not die of recurrent infections or SIRS but the same causes (cardiovascular, respiratory, cancer, etc.) as the control patients. The statistical associations are powerful, and the population studied is large.

There are some concerns with this study. Most of the patients (98.5%) were male. Co-morbidities were determined by diagnosis only and not stratified by severity. Other investigators have found significant problems with using diagnoses to reflect severity of co-morbid disease.

Since the same institutions were coding all patients, study and control, it was hoped that systematic coding errors would be avoided. However, it is possible that the SIRS patients were more carefully coded since they were the focus of another randomized study.

This study has identified an association between SIRS survival and increased mortality. As with all statistical methods, correlation cannot be proved to be the cause of the increased mortality. There may be some other unidentified factors that are responsible for both SIRS and reduced survival. Alternatively, there may be factors not identified in the control patient’s records that would change the predictive model and remove the association. The results of this study are provocative, their implications for patients are serious, and continued study of SIRS survivors is justified.