BACKGROUND AND DESIGN: Microbiological cultures of skin biopsy tissue are often recommended in immunocompromised patients with cancer and rashes, but in a previous study, they were usually sterile or grew clinically insignificant organisms. To examine the use and bacteriological results of these cultures more comprehensively, we reviewed records from all immunocompromised adults with cancer and acute rash on which skin biopsy was performed during 39 months on a bone marrow transplantation/acute leukemia unit of a university hospital (108 episodes of rash in 80 patients). RESULTS: Of the 158 cultures that were performed, one (1%; 95% confidence interval [CI], 0% to 4%) was a true positive; 11 (7%; 95% CI, 3% to 13%) were false positive; 143 (91%; 95% CI, 87% to 95%) were true negative; and three (2%; 95% CI, 1% to 6%) were false negative. Thus, the sensitivity of culture was 0.25, and the specificity was 0.93. Coagulase-negative Staphylococcus was the single pathogenic organism that grew, yet was judged to be a contaminant in three episodes. Among the 95 rashes in which fewer than four types of culture were performed, viral culture may have been helpful in one case (1%; 95% CI, 0% to 6%). CONCLUSIONS: Cultures of skin tissue from immunocompromised cancer patients with rashes were often unable to diagnose infection or the absence of infection. Clinical judgement was crucial to the interpretation of culture results.
BACKGROUND AND DESIGN: Microbiological cultures of skin biopsy tissue are often recommended in immunocompromised patients with cancer and rashes, but in a previous study, they were usually sterile or grew clinically insignificant organisms. To examine the use and bacteriological results of these cultures more comprehensively, we reviewed records from all immunocompromised adults with cancer and acute rash on which skin biopsy was performed during 39 months on a bone marrow transplantation/acute leukemia unit of a university hospital (108 episodes of rash in 80 patients). RESULTS: Of the 158 cultures that were performed, one (1%; 95% confidence interval [CI], 0% to 4%) was a true positive; 11 (7%; 95% CI, 3% to 13%) were false positive; 143 (91%; 95% CI, 87% to 95%) were true negative; and three (2%; 95% CI, 1% to 6%) were false negative. Thus, the sensitivity of culture was 0.25, and the specificity was 0.93. Coagulase-negative Staphylococcus was the single pathogenic organism that grew, yet was judged to be a contaminant in three episodes. Among the 95 rashes in which fewer than four types of culture were performed, viral culture may have been helpful in one case (1%; 95% CI, 0% to 6%). CONCLUSIONS: Cultures of skin tissue from immunocompromised cancerpatients with rashes were often unable to diagnose infection or the absence of infection. Clinical judgement was crucial to the interpretation of culture results.