Sw. Smith et al., Severe sepsis in the emergency department and its association with a complicated clinical course, ACAD EM MED, 5(12), 1998, pp. 1169-1176
Objective: Infection severity as determined by clinical criteria has been r
ecently classified and studied in hospitalized inpatients. The objective of
the study was to use modified criteria to determine the clinical course as
sociated with three levels of infection severity in infected patients admit
ted from the ED. Methods: This was a retrospective cohort study involving a
ll patients 18 years of age and older admitted through the ED of an urban t
eaching hospital during a four-month period whose primary reason for requir
ing hospitalization was an infection that was recognized in the ED. ED reco
rds were reviewed for criteria used to classify patients by three levels of
infection severity: no systemic inflammatory response syndrome, sepsis, an
d severe sepsis (SS). The relationships between these classifications as we
ll as certain clinical characteristics and any complicated clinical course
as measured by death and/or admission to an intensive care unit (ICU), and/
or prolonged hospitalization, were analyzed. Results: Of 408 patients enter
ed into the study, 138 (33.8%) fulfilled the criteria of SS in the ED. Pati
ents with SS in the ED had a mortality of only 4.3%, though with an increas
ed risk of dying compared with that of the other groups combined [relative
risk (RR) = 11.64, 95% confidence interval (CI) = 1.43 to 96.53], an increa
sed risk of ICU stay (RR = 7.65, 95% CI = 4.08 to 14.36), and an increased
risk of prolonged hospitalization (RR = 1.99, 95% CI = 1.38 to 2.88). Altho
ugh age over 60 years and several comorbid conditions also were identified
by univariate analysis as risk factors, multivariate analysis revealed that
only SS and diabetes' mellitus (DM) were independent predictors of a compl
icated course. In the authors' institution, the positive predictive value (
PPV) of SS for complicated clinical course was 0.48 and the negative predic
tive value (NPV) of no SS for no complicated course was 0.77. The PPV of [S
S + DM] was 0.83, and the NPV of [SS, DM, or both] was also 0.83. Conclusio
n: Although the strongest correlate of a complicated clinical course identi
fied in the ED is SS as defined by the study criteria, its specificity and
PPV are low. The mortality of ED patients with SS is much lower than the mo
rtality rates reported for inpatients with SS. SS as defined by the study c
riteria is too sensitive and therefore lacks utility in the ED setting.