285 EPIDEMIOLOGY OF MYCOBACTERIUM KANSASII DISEASE AMONG PATIENTS PRESENTING TO AN URBAN INNER-CITY HOSPITAL

BackgroundMycobacterium kansasii (MK) is an important cause of nontuberculous mycobacterial infection especially in the immunocompromised patient, but the clinical significance of positive cultures has been questioned in the past. Over the past decade, the number of patients at our institution with positive cultures for MK increased from < 10 per year (0.18/1,000 admissions) to over 25 per year (1.9/1,000 adm). Over this time period there was no change in the respiratory isolation policy to account for the increase. In October 2000 we instituted the use of an Mtb PCR assay on all smear-positive respiratory samples.PurposeTo examine the epidemiological, clinical, and radiologic features of M. kansasii infection over a 10-year period from 1994 to 2004.MethodsChart review of any patient with a culture positive for M. kansasii between 1994 and 2004.ResultsDuring the study period, 320 cultures grew MK from 178 patients. Culture sources for the 320 isolates were respiratory in 303 (94.5%) and 144 (47.5%) of these were AFB smear positive. Other sources included 9 (2.8%) blood and 1 each urine, thigh, stool, pancreas, and perirectal abscess. Among 178 MK patients, 133 (74.7%) were male, 165 (92.7%) were black, and median age was 39 yrs (range 22-79 yrs). Most (159 [89%]) were HIV positive and median CD4 was 19 cells/μL. Twelve (7%) of patients died prior to discharge. Of 102 patients with a single positive culture, 66 (65%) were smear-negative c/w 31 (41%) of 76 patients with multiple positive cultures (p < .005). Abnormal CXRs were found in 121 (72%) of 167 cases; 30 (18%) of these were consistent with PCP. For patients with abnormal CXRs and + smears, discharge TB or MK Rx was no less likely after use of Mtb PCR (13 of 30 prior to 10/2000 vs 8 of 21 after 10/2000).ConclusionM. kansasii infection increased at this institution over the last 10 years despite the advent of highly active antiretroviral therapy during this period. This infection predominantly affected severely immunocompromised patients as evidenced by the very low median CD4 count. Less than 50% of patients were discharged on TB or MK Rx at time of diagnosis, making it very unlikely that the majority of cases received the recommended 18-month course of therapy. This proportion was not affected by Mtb PCR testing. Mtb PCR testing may have affected length of stay. Long-term outcome for these patients is under review.