RESUMO
OBJECTIVE: To investigate an outbreak of healthcare-associated Burkholderia cepacia complex (BCC) primary bloodstream infections (BCC-BSI). DESIGN AND SETTING: Case-crossover study in a public hospital, a university hospital and a private hospital in Rio de Janeiro, Brazil, from March 2006 to May 2006. PATIENTS: Twenty-five patients with BCC-BSI. DESIGN: After determining the date BCC-BSI symptoms started for each patient, 3 time intervals of data collection were defined, each one with a duration of 3 days: the case period, starting just before BCC-BSI symptoms onset; the control period, starting 6 days before BCC-BSI symptoms onset; and the washout period, comprising the 3 days between the case period and the control period. Exposures evaluated were intravascular solutions and invasive devices and procedures. Potential risk factors were identified by using the McNemar chi(2) adjusted test. Cultures of samples of potentially contaminated solutions were performed. BCC strain typing was performed by pulsed-field gel electrophoresis using SpeI. RESULTS: The statistical analysis revealed that the use of bromopride and dipyrone was associated with BCC-BSI. A total of 21 clinical isolates from 17 (68%) of the 25 patients and an isolate obtained from the bromopride vial were available for strain typing. Six pulsotypes were detected. A predominant pulsotype (A) accounted for 11 isolates obtained from 11 patients (65%) in the 3 study hospitals. CONCLUSION: Our investigation, using a case-crossover design, of an outbreak of BCC-BSI infections concluded it was polyclonal but likely caused by infusion of contaminated bromopride. The epidemiological finding was validated by microbiological analysis. After recall of contaminated bromopride vials by the manufacturer, the outbreak was controlled.