285 Epidemiology and Risk Factors of Patients with Blood Stream Infection (BSI) due to Acinetobacter baumannii (AB) – A Multicenter Case-Control Study

Saturday, April 2, 2011
Trinity Ballroom (Hilton Anatole)
Teena Chopra, MD , Detroit Medical Center, Wayne State University, Detroit, MI
Paul C. Johnson, MD , Detroit Medical Center, Wayne State University, Detroit, MI
Hardik Doshi, MBBS , Detroit Medical Center, Wayne State University, Detroit, MI
Jing J. Zhao, PharmD , Detroit Medical Center, Wayne State University, Detroit, MI
Jason M. Pogue, PharmD , Detroit Medical Center, Wayne State University, Detroit, MI
Kayoko Hayakawa, MD, PhD , Detroit Medical Center, Wayne State University, Detroit, MI
Odaliz Abreu-Lanfranco, MD , Detroit Medical Center, Wayne State University, Detroit, MI
Sapna Parmar , Detroit Medical Center, Wayne State University, Detroit, MI
Dror Marchaim, MD , Detroit Medical Center, Wayne State University, Detroit, MI
Sorabh Dhar, MD , Detroit Medical Center, Wayne State University, Detroit, MI
Keith S. Kaye, MD, MPH , Detroit Medical Center, Wayne State University, Detroit, MI

Background: AB is a hospital pathogen that has become increasingly common in hospitals. Scant data exist regarding the factors predictive of BSI due to AB in hospitalized patients. 

Objective: The objective of this study was to analyze risk factors for BSI due to AB. 

Methods: A retrospective matched case-control study including 245 episodes of AB BSI and 245 matched controls without infection. The study was conducted at the Detroit Medical Center and included patients cared for from January 2006 to April 2009.  Patients with positive blood cultures for AB were identified from laboratory data.  Controls were matched to case patients in a 1:1 ratio based on admitting hospital, admitting unit, and at-risk period (time admission to positive culture date for cases).  Patient variables collected included demographics, comorbid conditions, antibiotic exposure, surgeries, mechanical ventilation, central lines, Foley catheters, and inpatient mortality. Antibiotic exposure in cases was defined as administration of an antibiotic from thirty days prior to culture date up to one day prior to culture date.  In controls, antibiotic exposure was defined as administration of any antibiotic from thirty days prior to discharge up until discharge.  Logistic regression was used to identify independent risk factors for BSI due to AB.

Results: Mean age of the case and control cohorts was 55±24 and 55±24 (p=0.6).  47% of cases and 52% of controls were male (p=0.23). Seventy-three % of cases and 81% of controls were African-American (p=0.04).  The median Charlson's comorbidity index was greater in cases (4, IQR 2-8) than in controls (0-4) (p<0.0001).  60% of cases were admitted from home compared to 90% of controls (p<0.0001).  

The following variables, excluding antibiotics, were statistically significant in multivariate analysis (Table).

 

            Risk Factors

Odds Ratio (95% CI)

p-value

Central Line Placement

6.1 (2.68-14.22)

<0.0001

Foley Catheter Placement

0.4 (0.20-0.88)

0.02

Serum Albumin < 2.5 g/dL

0.1 (0.06-0.26)

<0.0001

When antibiotics were added to the model, the following were associated with aquistion of BSI due to AB: carbapenems (OR 5.3 [2.21-12.56], p=0.0002), colistimethate (OR 3.9 [1.08-14.24], p=0.04) and use of β-lactams including β-lactams/β-lactamase-inhibitor (BLABLI) (OR 1.7 [0.99-2.86], p=0.05).

Conclusions : Exposure to various broad spectrum antimicrobials, including carbapenems, colistimethate, and β-lactams/BLABLIs were each strongly associated with acquisition of BSI due to AB.