Abstract
The objective of the current study was to update parameterization of mathematical simulation models for foot-and-mouth disease (FMD) spread in cattle utilizing recent knowledge of FMD virus (FMDV) pathogenesis and infection dynamics to estimate the duration of distinct phases of FMD. Specifically, the durations of incubation, latent, and infectious periods were estimated for 3 serotypes (O, Asia1, and A) of FMDV, individually and collectively (pan-serotypic). Animal-level data were used in Accelerated Failure Time (AFT) models to estimate the duration of the defined phases of infection, while also investigating the influence of factors related to the experimental design (exposure methods) and virus serotype on disease progression. Substantial influences upon the estimated duration of distinct phases of FMD included the quantity of viral shedding used as a proxy for the onset of infectiousness, virus serotypes, and experimental exposure methods. The use of detection of any viral RNA in nasal secretions as a proxy of infectiousness lengthened the total infectious period compared to use of threshold-based detection. Additionally, the experimental system used to infect the animals also had significant effects on the duration of distinct phases of disease. Overall, the mean [95% Confidence Interval (CI)] durations of pan-serotype disease phases in cattle were estimated to be: incubation phase = 3.6 days (2.7-4.8), latent phase = 1.5 days (1.1-2.1), subclinical infectious phase = 2.2 days (1.5-3.5), clinical infectious phase = 8.5 days (6.2-11.6), and total infectious phase = 10.8 days (8.2-14.2). This study highlights the importance of identifying appropriate proxy measures to define the onset and duration of infectiousness in FMDV-infected cattle in the absence of actual transmission data. Additionally, it is demonstrated herein that factors associated with experimental design, such as virus exposure methods, may significantly affect disease progression in individual animals and should be considered when data is extrapolated from experimental studies. Given limitations in experimental data availability, pan-serotypic parameters which include all routes of exposure and a threshold-defined onset of infectiousness may be the most robust parameters for exploratory disease spread modeling approaches, when information on the specific virus of interest is not available.