Molecular epidemiology of tuberculosis in a sentinel surveillance population

Abstract

We conducted a population-based study to assess demographic and risk-factor correlates for the most frequently occurring Mycobacterium tuberculosis genotypes from tuberculosis (TB) patients. The study included all incident, culture-positive TB patients from seven sentinel surveillance sites in the United States from 1996 to 2000. M. tuberculosis isolates were genotyped by IS6110-based restriction fragment length polymorphism and spoligotyping. Genotyping was available for 90% of 11923 TB patients. Overall, 48% of cases had isolates that matched those from another patient, including 64% of U.S.-born and 35% of foreign-born patients. By logistic regression analysis, risk factors for clustering of genotypes were being male, U.S.-born, black, homeless, and infected with HIV; having pulmonary disease with cavitations on chest radiograph and a sputum smear with acid-fast bacilli; and excessive drug or alcohol use. Molecular characterization of TB isolates permitted risk correlates for clusters and specific genotypes to be described and provided information regarding cluster dynamics over time.

Barbara A. Ellis, Guest Editor. Dr. Ellis is a senior microbiologist with the National Center for Infectious Diseases, Centers for Disease Control and Prevention. Her research interests focus on the molecular epidemiology of infectious diseases, rodent-borne zoonotic diseases, and bioterrorism preparedness. Her work has included disease ecology studies of rodent-borne hemorrhagic fever viruses, molecular characterization of novel bartonellae, and molecular epidemiologic studies of Mycobacterium tuberculosis.

Figure 1

Numbers of tuberculosis cases, cumulative proportion of cases with isolates in genetic clusters, and maximum genetic cluster size from seven sentinel surveillance sites by quarter that verified case was counted, 1996–2000. Numbers of cases with isolates that had unique genotypes and those with isolates that were in genetic clusters are shown separately.

Figure 2

Average annual incidence of tuberculosis for seven sentinel surveillance sites and percentage of cases with isolates in genetic clusters, 1996 to 2000. Spearman correlation coefficient and probability of correlation between incidence and percentage of cases clustered are given.

Figure 3

Number of cases with isolates that had unique genotypes (“not clustered”) and those in genetic clusters for U.S.-born (A) and foreign-born persons (B) by number of copies of IS6110.