A matched case-control study was conducted at the University of Michigan Health System from January 1, 2000 through September 30, 2008. Cases (n=299) were adults who had both a positive urine and blood culture with the same microorganism during hospitalization. Patients with a positive urine culture within the first 2 days of admission or who were admitted for reason of septicemia or bacteremia were excluded. Controls (n=670) were selected by incidence density sampling and were a sample of adults with a positive urine culture who were at risk of a bloodstream infection but did not develop such an infection. Controls were matched to each case by calendar time (within 120 days) when the bloodstream infection occurred in the case. At maximum, there were 4 controls per case but, owing to constraints of the control definition, some cases were matched to 1 to 3 controls. A similar period was compared for cases and controls. That is, if the urinary tract-related bloodstream infection occurred on the 20th day after hospital admission for the case, a similar time period was evaluated in the matched control (from admission to the 20th day of hospitalization). For purposes of this investigation, this day was labeled the “index date.”
Information was extracted from electronic medical records, with additional medical chart review by an infectious diseases physician. Conditional logistic regression was used, which accounted for the matched design. Adjusted models included RBC, platelets, fresh-frozen plasma, medication use during the 2 days prior to the index date (antibacterial, antifungal, antimicrobial, antiviral, statin, insulin, and immunosuppressants), age, race, sex, surgery (cardiovascular, gastrointestinal, other), cardiovascular disease, cancer and diabetes mellitus. To evaluate possible nonlinear associations between the volume of blood component and the outcome, fractional polynomials were used. For those analyses in which the age of the RBC unit was evaluated, some patients received multiple units and therefore, we used the expiration date for the oldest RBC unit given.
This study was approved by the University of Michigan Health System institutional review board.