CROSS-CONTAMINATION OF SPECIMENS WITH MYCOBACTERIUM-TUBERCULOSIS - CLINICAL-SIGNIFICANCE, CAUSES, AND PREVENTION

At the Veterans Affairs Lakeside Medical Center, two episodes of speci men cross-contamination with Mycobacterium tuberculosis were detected during a 54-month period by molecular strain typing using DNA restrict ion fragment length polymorphism for 3 patients without clinical or ra diologic signs of tuberculosis (TB). A cross-contaminated specimen was the only culture-positive specimen for each of the 3 patients. Labora tory features of cross-contamination included acid-fast smear negativi ty, growth only in broth or solid medium, and growth in solid medium w ith 5 or fewer colonies. Retrospective analysis demonstrated identical features for occasional culture-positive specimens from 54 patients w ith TB during the same period. However, productive cough, pleural pain , weight loss, night sweats, chest radiograph results suggestive of TB , positive tuberculin skin testing, and/or multiple culture-positive s pecimens were invariably present in patients with TB with such specime ns. Most patients with TB (50/54; 93%) had multiple specimens positive in culture for M tuberculosis, and the few patients with TB with sing le culture-positive specimens were symptomatic. These results indicate that correlation with clinical manifestations is necessary to determi ne the significance of isolated, acid-fast smear negative, and/or low- yield culture-positive specimens. Although the prevalence of specimen cross-contamination is low (0.1%), possible sources (especially the us e of single-reagent delivery systems for multiple specimens) should be eliminated by mycobacteriology laboratories.