Hospital-associated Infection is Associated with Longer Post-infection Length of Stay (LOS) among Pediatric Patients with S. aureus Bacteremia

Background: S. aureus infections among children are thought to be predominantly community-associated. However, limited data exist on the acquisition of pediatric infections and associated outcomes. Post-infection LOS is an important outcome that has major clinical and economic consequences for both patients and healthcare systems.

Objective: We aimed to compare patient characteristics and post-infection LOS among children with community-associated vs. hospital-associated S. aureus bacteremia (SAB).

Methods: This retrospective cohort study included all pediatric patients (<18 years) with SAB admitted to a tertiary-care facility from 1/1/2002 to 6/30/2007. Healthcare-associated hospital-onset (HA-HO) infection was defined as culture collection >48 hours of admission. Healthcare-associated community onset (HA-CO) infection was defined as culture collection ≤48 hours of admission but had ≥1 of the following prior healthcare exposures: admission in past year, malignancy, hemodialysis or prior history of MRSA. Community-associated community-onset (CA-CO) infection was defined as culture collection ≤48 hours of admission without prior healthcare exposures. Post-infection LOS was defined as time from culture collection to discharge. The Wilcoxon rank sum test was used for bivariate comparisons. Adjusted Hazard Ratios (HR) and 95% Confidence Intervals (CI) were calculated using Cox proportional hazard models.

Results: Among 98 cases of pediatric SAB, 23% were CA-CO, 17% were HA-CO and 60% were HA-HO. Patients with HA-HO were significantly younger compared to both CA-CO (p=0.01) and HA-CO patients (p<0.01) (Table). Seven patients died within 30 days of culture collection (3 had CA-CO infection, 3 had HA-HO infection and 1 had HA-CO infection; p=0.4). CA-CO and HA-CO infections had similar LOS (p=0.45) but both were associated with shorter post-infection LOS compared to HA-HO infections (p<0.01 for both). After controlling for age, methicillin resistance, intensive care unit admission before culture collection and time to receipt of appropriate therapy, both CA-CO and HA-CO infections were associated with shorter LOS compared to HA-HO infections (CA-CO HR: 0.5, 95% CI: 0.28, 0.91; HA-CO HR: 0.33, 95% CI: 0.18, 0.63).

Conclusions: Pediatric patients with community-associated infections, including those with healthcare exposures, had shorter post-infection LOS compared to patients that acquired SAB during their index hospitalization. Although much attention has been given to pediatric CA-CO infections recently, we must still continue to make efforts to prevent pediatric HA-HO infections given their high clinical and economic impact.

Table:

	CA-CO SAB	HA-CO SAB	HA-HO SAB
Median Age, years (Interquartile range [IQR])	4 (0, 13)	2 (1, 14)	0 (0, 10)
Methicillin Resistant	31%	35%	45%
Median Post-infection LOS, days (IQR)	14 (5, 26)	7 (4, 11)	32 (15, 63)