Subject: CWD Modeling Paper that appeared in the Journal of Wildlife Management, April 2001

In case you did not wade through the 37 pages in the CWD Modeling paper here are some excerpts from it. The entire paper can be viewed at:

http://www.nrel.colostate.edu/projects/cwd/papers/gross.pdf

DS note: We know from disinfections and leaving research pens empty for one year that environmental contamination is a method of transferring CWD. Therefore this model is a best-case scenario. Environmental contamination effectively deletes the chance occurrences of the infected animals dying off before transmission occurs as their carcasses remain infectious in the environment after death.

CHRONIC WASTING DISEASE IN MULE DEER: DISEASE DYNAMICS AND CONTROL

John E. Gross, Natural Resource Ecology Laboratory, Colorado State University, Fort Collins, CO 80523, USA
Michael W. Miller,3 Colorado Division of Wildlife, Wildlife Research Center, 3 17 West Prospect Road, Fort Collins, Colorado 80526, USA

We developed a mechanistic model to simulate dynamics of chronic wasting disease (CWD) in mule deer. Parameters were estimated from observations of infected and uninfected mule deer in Colorado.

Disease was not sustained in projected populations when transmission rates were low, but CWD eliminated populations when more realistic transmission rates were used.

Simulated selective culling programs revealed the importance of initiating control while CWD prevalence was low (~0.01) (1%). Low selective culling rates (~20% of infected populations) effectively eliminated CWD if initiated when prevalence was low, but the likelihood of control diminished rapidly as prevalence increased. Management programs will likely require an effort sustained over many decades if eliminating CWD is the desired goal.

Once introduced and established, most infectious diseases are extremely difficult to eliminate from free-ranging populations.

The CWD agent is presumed to be shed in some combination of saliva, feces, urine, and/or placental tissues and fluids.

We believe infected deer become progressively "infectious" as disease progresses.

Whether residual excreta or decaying carcasses can serve as additional environmental sources of CWD infection has not been determined (Williams and Young 1992, Miller et al. 1998, 2000; M. W. Miller, unpublished data), and consequently is not considered in our model.

We assumed that each infectious animal produced an estimable number of infectious contacts per unit time, and that all individuals in the host population had an equal probability of contacting an infectious dose.

To evaluate each set of parameter values we began each simulation with a population of 1,000 deer that included 4 infectious and 4 latent 2-year-old females. For each set of parameters, we conducted 250 simulations that each lasted 100 years.

In contrast, population growth rates were highly sensitive to changes in survival of adult females. Mean population size and productivity of simulated populations infected with CWD contrasted sharply with uninfected populations.

Once CWD was firmly established in the population (e.g., prevalence increased to about 2%), the proportion of infectious animals in the population uniformly and rapidly increased and populations declined at a relatively consistent rate (Fig. 3)

Model results closely matched independent field observations of age-specific prevalence rates and the ratio of latent to infectious deer (Fig. 4).

Growth and productivity of infected populations were profoundly influenced by even low rates of infection (Fig. 5).

Under these conditions, CWD persisted to year 100 in only 2 of 250 simulations. The extinction of most infected populations resulted in highly variable rates of prevalence after about year 60, particularly in simulations with a high transmission rate (i.e., > 0.7 contacts per period; Fig. 6B).

Only selective culling treatments that eliminated fewer than 0-20% of infected (latent and infectious) deer failed to eliminate the disease from virtually all populations after 80 years of treatment (Fig. 7).

There was a dramatic decline in the effectiveness of a specific selective culling strategy to eliminate CWD when the program was initiated at a prevalence of 0.05 (5%) compared to 0.01 (1%).

Our simulations suggest that' efforts to prevent spread by reducing dispersal will require that the number of animals dispersing from infected populations must remain very low.

To the extent that modeled mechanisms of CWD transmission appear to offer at least a reasonable approximation of disease processes occurring in nature, it follows that this model provides plausible forecasts of future epidemic trends. The projected trends are, to say the least, unsettling. Left unmanaged, our model forecasts 2- to 4-fold increases in CWD prevalence over the next several decades with disease abating only with the extinction of infected deer populations. In light of an observed lo-fold increase in CWD cases among captive mule deer over a 4-year period (M. W. Miller, unpublished data), such trends are clearly plausible.

CWD was naturally eliminated from simulated populations only under 1 of 2 conditions...either a low transmission rate or to the death(s) of the few infectious individuals before transmission occurred...when the few infectious individuals remaining died before transmitting CWD, simply through chance events.

The slow initial rate of increase in prevalence of CWD and the eventual decimation of infected populations suggest that long-term persistence of CWD is likely to occur as a result of dispersal and the spatial structure of deer populations. Heavily infected populations may die out, while recently infected populations will continue to provide a source of infection and lead to long-term persistence and geographic spread of CWD.

In northeastern Colorado, radiomarked deer typically have a high fidelity to seasonal ranges, but ranges used in summer and winter may be separated by ~25 km dispersal distances for fawns may exceed 100 km.

In particular, the potential role of environmental contamination in CWD epidemiology warrants further exploration (Miller et al. 1998, 2000).

Because the present model included no spatial attributes, potential for and patterns of geographic spread cannot be assessed adequately here.

As presently applied (e.g., removing clinical cases when noticed or reported by the public), however, selective culling is probably not a particularly effective strategy for CWD management. Unfortunately, there are 2 formidable obstacles to large-scale institution of more effective programs in the foreseeable future. First, practical and reliable tests for detecting preclinical or subclinical CWD under field conditions are lacking; as a further constraint, all antemortem tests currently showing promise require laboratory processing of samples before results are known.

In light of present technological obstacles to detection of both infected animals and infected populations, delays in management intervention seem inevitable. The consequences of such delays, however, may be equally severe. Our model forecasts that allowing prevalence to increase to 0.05 (5%) before intervening doubles the time required to have a 50:50 chance of eliminating CWD from the infected population. At prevalence rates currently seen in some affected deer management units in Colorado and Wyoming (Miller et al. 2000), model projections suggest CWD may be essentially unmanageable without substantial investment in selective culling applied to a large proportion of those populations in the very near future.