LETTER

Top Letter References

Recent developments in the automation of blood culture systems have greatly decreased the times required for detection of many microorganisms. This has been especially useful in the rapid detection of slow-growing bacteria. Recent reports have described more rapid detection under both in vivo and in vitro conditions for members of the genus Brucella. Casas et al. (2) studied the suitability of the Bact/Alert system (Organon Teknica, Durham, N.H.) for the in vitro detection of Brucella spp. and reported a mean detection time of 2 to 3 days. The same authors also studied Brucella growth in vivo by sequential sampling in five patients. The Bact/Alert system detected Brucella melitensis in only one bottle after 2.9 days of incubation, while detection was delayed for up to 10 to 20 days by a blind pass in the remaining seven bottles.

More recently, Ruiz et al. (3) evaluated the performance of three blood culture systems, Hémoline diphasic medium (bioMeriéux, Marcy l'Etoile, France), Bactec Plus Aerobic/F (Becton Dickinson, Paramus, N.J.), and Vital Aer (bioMeriéux), in the diagnosis of 17 cases of brucellosis. The mean detection times obtained were 136.94, 92.37, and 162.18 h, respectively, with the earliest positive detection made with the Bactec system after 59 h.

We have been using the Bact/Alert system in our laboratory for routine blood cultures over the past 2.5 years. This system was applied to blood samples obtained from patients with and without suspicion of brucellosis. Samples for three complete sets (one aerobic and one anaerobic culture bottle each) were taken at 30-min intervals. In suspected cases of brucellosis, bottles were maintained up to 30 days and then subcultured unless any sign of positivity was detected. Over a recent 8-month period we have isolated five Brucella strains from blood cultures of five patients. Patient 1 (diagnosis of suspected brucellosis) had a primary aerobic culture bottle that was positive after 64.8 h of incubation and a second aerobic bottle that was positive after incubation for 86.4 h. For patient 2 (diagnosis of endocarditis), all three aerobic bottles were positive within 44, 46.3, and 48 h, respectively. For patient 3 (diagnosis of pancreatic abscess), two aerobic culture bottles were positive after incubation for 86.4 and 88.4 h, respectively. In addition, a culture bottle inoculated with a sample of pancreatic abscess fluid became positive after only 13.3 h. Patient 4 (diagnosis of febrile syndrome) and patient 5 (diagnosis of pyelonephritis) each had one aerobic bottle that was positive within 72 h. All the strains were identified as Brucella spp., and serum samples from all five patients were shown to be reactive to Brucella by standard serological tests.

In summary, our data obtained by using the Bact/Alert system indicate a more rapid detection of Brucella in vivo than has been shown in previous reports, with our mean time of detection being 67.8 h. Laboratories should be aware of the potential risk when these strains are isolated and the possibility of misidentifying this bacterium as Moraxella phenylpyruvica as has been described elsewhere (1). We observed this using the API 20NE bacterial identification system (bioMérieux). Finally, it is advisable to suspect the presence of Brucella spp. when positivity is detected only in aerobic bottles after incubation for 2 to 3 days, Gram smears from culture bottles are difficult to resolve, and tiny colonies are observed at 24 h after subculture to solid media.