Human Plague Infection | CDC EID

EID Journal Home > Volume 15, Number 10–October 2009

Volume 15, Number 10–October 2009
Dispatch
Fine-scale Identification of the Most Likely Source of a Human Plague Infection
Rebecca E. Colman, Amy J. Vogler, Jennifer L. Lowell, Kenneth L. Gage, Christina Morway, Pamela J. Reynolds, Paul Ettestad, Paul Keim, Michael Y. Kosoy, and David M. Wagner
Author affiliations: Northern Arizona University, Flagstaff, Arizona, USA (R.E. Colman, A.J. Vogler, P. Keim, D. Wagner); Centers for Disease Control and Prevention, Fort Collins, Colorado, USA (J.L. Lowell, K.L. Gage, C. Morway, M.Y. Kosoy); and New Mexico Department of Health, Santa Fe, New Mexico, USA (P.J. Reynolds, P. Ettestad)

Suggested citation for this article

Abstract
We describe an analytic approach to provide fine-scale discrimination among multiple infection source hypotheses. This approach uses mutation-rate data for rapidly evolving multiple locus variable-number tandem repeat loci in probabilistic models to identify the most likely source. We illustrate the utility of this approach using data from a North American human plague investigation.

Linking human disease events to likely sources of infection has been advanced by molecular epidemiology. However, isolates from several potential infection sources often are similar, and none may exactly match the clinical isolate genotype, especially if the methods used provide high discrimination (1). Conclusions from partial-match genotypes are problematic but may provide the only data for weighing the relative importance of similar source genotypes. Even perfect-match genotypes do not preclude partial-match sources as likely infection sources (2). We present a probabilistic approach based on mutation rates that can be used to identify the most likely source of infection. Our example is human plague, but the approach could be applied to other diseases for which data on marker mutation rates are available (3).

Plague is caused by the bacterium Yersinia pestis. Because Y. pestis is an obligate pathogen that continuously cycles between rodents and fleas, mutations are generated regularly and can be observed among even closely related isolates (1). Human contact with infected fleas or rodents can result in human plague (4). Plague is rare in the United States, with <20 cases in 2006 (5) but is of concern because of the potential use of Y. pestis as a biological weapon (6). Thus, the ability to link a human plague isolate to a likely source has implications for investigating both natural disease and bioterrorism events.

Multiple locus variable-number tandem repeat (VNTR) analysis (MLVA) is useful for molecular epidemiologic studies of Y. pestis because of its discrimination power (1,7,8). We previously used MLVA to genotype the human isolate described below and queried the resulting genotype against a database containing genotypes from hundreds of Y. pestis isolates (9). This statistical approach identified isolates that most closely matched the human isolate and confirmed its most likely coarse geographic origin (northern New Mexico). However, this set of near matches from the database query included isolates representing several different potential local infection sources, leaving the most likely fine-scale source unclear. The human and environmental isolates were indistinguishable with pulsed-field gel electrophoresis (PFGE); thus, the most likely fine-scale source could not be identified (10).

abrir aquí para acceder al documento CDC completo, del cual se reproduce una mínima parte:
Human Plague Infection | CDC EID