Social Network Analysis and Whole Genome Sequencing Aid in TB Outbreak Investigation
By Amesh A. Adalja, MD, March 4, 2011
Tuberculosis (TB) outbreak investigation traditionally has been conducted through “shoe-leather epidemiology,” in which investigators start with an index case and work backward through that person’s contacts, testing each for TB in order to identify other cases and, eventually, the source. Researchers from British Columbia have just published results of a TB outbreak investigation in which the combined use of whole-genome sequencing and social-network analysis produced results not likely to have been achieved through traditional methods. This study suggests that new approaches and technologies have the potential to advance and enhance contact tracing for TB and other contagious infectious diseases.
Investigation of a 41-Case TB Outbreak in British Columbia
In May 2006, a case of extra-pulmonary TB was diagnosed in a community in British Columbia. Two months later, a case in an infant prompted a public health investigation. By the end of October 2006, 41 cases had been identified, 37 of which were laboratory confirmed. Mycobacterial interspersed repetitive unit-variable-number tandem repeats (MIRU-VNTRs) analysis determined that Mycobacteria tuberculosis isolates from all cases were identical.
Standard contact tracing was used to characterize the outbreak, but when it did not point to a clear source, the investigators initiated a social-network analysis. The first step entailed characterizing the social network of case patients, which was accomplished through a questionnaire designed to pinpoint shared social settings in which TB may have been transmitted. The questionnaire, which was administered to 9 case patients, elicited information about drug and alcohol use, residential and travel history, places of social aggregation, and social networks.
Social-network analysis led investigators to the most likely source case—an adult with smear positive cavitary disease who had been symptomatic for at least 8 months prior to detection of the index case. The source case was connected to almost all of the early cases through either direct contact or shared social locales. The social-network analysis also led to identification of previously unreported interactions and shared locations, including 2 hotels, 2 community centers, and several crack houses.
Whole Genome Sequencing Identifies 2 Outbreaks
Although MIRU-VNTR typing indicated that the M. tuberculosis isolates were identical, whole genome sequencing revealed 2 distinct lineages, signaling the occurrence of 2 simultaneous outbreaks that shared risk factors and transmission dynamics. When combined with the social-network analysis, the nature of the outbreaks came into even sharper focus, leading to, for instance, identification of a super spreader who played a crucial role in 1 of the 2 outbreaks.
New Approaches to Augment Traditional Contact Tracing
Standard contact tracing has long been effective in investigations of disease outbreaks caused by direct contact. However, the results of this investigation indicate that traditional methods may not be adequate for identifying cases acquired through airborne transmission, such as TB, because a source patient is not likely to be able to identify all of the people with whom they've shared a location. In this study, social-network analysis revealed connections not otherwise discovered. While this type of analysis is not new—traditional epidemiologic investigation typically involves inquiries into hobbies, workplaces, social venues, and the like—the methods employed in this study were designed to elicit a social networking profile for each of the case patients queried.
The other crucial component of this investigation—whole genome sequencing—also has the potential to aid in outbreak investigation. Indeed, it was only after the precise genetic constitution of the microorganism was identified that it became clear that 2 outbreaks had occurred simultaneously in British Columbia, a fact that otherwise may have been lost entirely.
Together, the 2 tools enabled a more complete and nuanced understanding of this outbreak while also leading to identification of a previously unsuspected source. These tools appear to hold promise for future outbreak investigations.
Gardy JL, Johnston JC, Ho Sui SJ, et al. Whole-genome sequencing and social-network analysis of a tuberculosis outbreak. N Engl J Med 2011;364:730-739.