Comparison ofLysis-Centrifugation withLysis-Filtration anda Conventional Unvented Bottle forBloodCultures

systemhadthe advantage ofhaving colonies immediately available forfurther testing. Thecontamination ratewiththelysiscentrifugation system was 3%,compared with6% withthelysis-filtration system and0.4%withbrainheart infusion. Thecritical importance ofbloodcultures totheproper diagnosis andmanagementofpatients hasbeenthestimulus forcontinuous efforts byclinical microbiologists todevelop methods which are more sensitive andrapid. Muchofthis workhasbeentodefine theoptimal media, blood-to-medium ratios, andtypesofanticoagulants andadditives, aswell as appropriate blind subculture or smear times. Development ofaradiometric detection systemprovided asemiautomated procedure (BACTEC;Johnston Laboratories, Towson, Md.)that proved advantageous formany laboratories. More recently, thedevelopment ofa lysis-centrifugation blood culture system(Isolator; DuPontCo.,Wilmington, Del.) by DornandSmith(2)hasstimulated considerable interest. Thisprocedure hasseveral advantages inthat itconcentrates andthenseparates microorganisms fromtheplasma and thereby fromtheantibiotics andother antibacterial factors that might also bepresent. Theorganisms aretheninoculateddirectly ontoplate media, whichallows formore rapid availability ofmicrobial colonies andalsopermits counting ofthenumberofcolonies present. Culture ofbloodby a methodoflysis-filtration was developed inourlaboratory (4,12,14,15)tofit theneedsof ourpatient population, whichconsisted ofmany oncology patients whowere on awidevariety aswell aslarge dosages ofantimicrobial agents. Results fromearlier workwiththis systemsuggested that itwas more sensitive thanconventionalsetsofaerobic andanaerobic bottles fora variety of organisms tested (15). Itwas ofinterest tous tocomparethe sensitivities ofthelysis-centrifugation andlysis-filtration blood culture systems. Anunvented conventional bottle was also included toprovide optimal anaerobic coverageandalso topermit some comparison witha standard method.