Surveillance cultures in bone marrow transplant recipients: worthwhile or wasteful?

A prospective study of the value of surveillance cultures was performed in a bone marrow transplant (BMT) unit among 48 consecutive patients. All patients were admitted to laminar airflow or high-efficiency particulate air (HEPA) filtered rooms, maintained on reduced microbial diets and received oral non-absorbable antibiotics. With the onset of neutropenia, all patients received imipenem/cilastatin and 17 patients received low-dose amphotericen B 0.1 mg/kg/day. Pre-transplant and weekly post-transplant cultures of the stool, throat and urine were obtained on all patients. Nasal and vaginal cultures were performed on 26 patients. Sixteen patients developed 23 documented infections. The sensitivity of surveillance cultures for all infections was 38%, specificity 25%, positive predictive value 20% and negative predictive value 44%. When stratified by organisms, the sensitivity, specificity, positive predictive value and negative value were: Gram positive infections, 33%, 36%, 11%, 70%, Gram negative infections, 17%, 88%, 17%, 88%; fungal infections 37%, 50%, 11%, 75%; and Candida albicans, 100%, 57%, 14%, 100%. These data suggest that surveillance cultures may be useful to exclude C. albicans infections but are of limited value in predicting other types of infections in recipients of BMT.