13

13

Disease Patterns in the CWD Eradication Zone

Final Report

Prepared by: Dr. Michael D. Samuel

December 2006

14

14

Project Summary: This project investigated the infection patterns of chronic wasting disease (CWD) in the Disease Eradication Zone (DEZ) in south-central Wisconsin. Infection patterns related to age and sex, vulnerability of infected animals to harvest, and factors related to transmission of CWD among adult females were determined in this study. Findings in the research project are summarized as follows. Chronic wasting disease (CWD) is a fatal disease of white-tailed deer (Odocoileus virginianus) caused by transmissible protease resistant prions. Since the discovery of CWD in southern Wisconsin in 2001, more than 20,000 deer have been removed from a >2,500 km2 disease eradication zone surrounding the three initial cases. Nearly all deer removed were tested for CWD infection and sex, age, and harvest location were recorded. Our analysis used data from a 310 km2 core study area where disease prevalence was higher than surrounding areas. We found no difference in harvest rates between CWD infected and non-infected deer. Our results show that the probability of infection increased with age and that adult males were more likely to be infected than adult females. Six fawns tested positive for CWD, five fawns from the core study area, including the youngest (5 months) free-ranging cervid to test positive. The increase in male prevalence with age is nearly twice the increase found in females. We concluded that CWD is not randomly distributed among deer and that differential transmission among sex and age classes is likely driving the observed patterns in disease prevalence. We discuss alternative hypotheses for CWD transmission and spread and, in addition, discuss several possible non-linear relationships between prevalence and age. Understanding CWD transmission in free-ranging cervid populations will be essential to the development of strategies to manage this disease in areas where CWD is found as well as for surveillance strategies in areas where CWD threatens to spread. We evaluated the hypothesis of socially facilitated transmission of chronic wasting disease (CWD) among adult female white-tailed deer using spatial location and genetic relatedness for 1,387 female deer, as well as the spatial location of 1,321 adult male deer harvested during 2002-2004 CWD control efforts in Wisconsin, USA. There was little correlation between genetic relatedness and geographic distance among all pairs of adult females separated by up to 10 km. At small distances (<0.8 km), CWD positive deer were more related than random individuals in the population, indicating a weak association between relatedness and CWD infection. This relationship was confounded by a low degree of spatial aggregation of related females compared to previous theories. Infection in adult females was strongly influenced by closely related females (full-sibling, mother-offspring) that were spatially proximate (<3.2 km). To a lesser extent, infection was also influenced by the number of infected nearby females (<3.2 km). In contrast, infection was not influenced by less genetically related females (i.e. half-sibling, grandmother-granddaughter, cousin) that were also expected to be in the same social group. Our results suggest a hierarchy of CWD transmission within social groups based on familial relationships. Our results indicate that direct (deer-to-deer) transmission of CWD likely occurs between closely related female deer. CWD transmission also appears to occur among spatially proximate females. However, we cannot determine whether occasional direct contact or contact with a contaminated environment is responsible for increased infection among proximate females. It is likely that direct and environmental (deer-environment-deer) transmission occurs in this epidemic. The influence of spatially proximate females and close female kin on CWD infection does not extend beyond a 2-3 km radius. In addition, the spatial relationship between

15

15

infected females and among infected males and females suggests that CWD transmission operates on a local scale of 20-30 km2. Project Background/Justification: Chronic Wasting Disease (CWD) belongs to the family of diseases known as transmissible spongiform encephalopathies (TSEs). TSEs include such diseases as scrapie in sheep, bovine spongiform encephalopathy (BSE) in cattle (aka Mad Cow Disease) and Creutzfeldt-Jakob disease of humans. TSEs are diseases of the nervous system that result in distinctive lesions in the brain. The causative agent of CWD is not known; however, the disease is associated with a modified protein (prion). These modified proteins are typically found in nervous and lymphatic tissues.

CWD infects elk, white-tailed deer, and mule deer, but is not known to infect livestock or humans. No treatment is known and the disease is typically fatal. Infected deer and elk can appear robust and healthy in the early stages of CWD and may take several years before they show clinical signs of the disease. The clinical signs are not unique to the disease and can be due to other conditions such as malnutrition. Currently all testing for CWD requires the microscopic examination of a specific portion of the brain or lymphoid tissue. Recently, a biopsy technique for tonsilar tissues from live deer has been developed; however, this test only works for white-tailed deer and mule deer but not for elk. The mode of transmission of CWD between animals is not known, although direct contact between infected and non-infected animals via exposure to excreta (e.g., saliva, urine, feces) is the most likely route of transmission. Contamination of soil by excreta from infected animals is thought to be another route of transmission, particularly among captive herds of deer and elk. However, the role of environmental contamination in free-ranging animals is not clearly understood.

The spread of CWD in wild animals is of great concern. The disease was originally described in captive animals 35 years ago in Colorado. However, over the last five years, CWD has been detected in wild cervids in several surrounding states and Canada. In early 2002, CWD was reported in wild deer in South Dakota, Wisconsin, and now in New Mexico. The recent detection of CWD in the wild white-tailed deer herd in Wisconsin is of particular concern. White-tailed deer appear more susceptible than mule deer and elk to CWD with a greater percentage of the herd becoming infected. Until now, CWD was found in white-tailed deer herds in Colorado, Wyoming, and Nebraska where deer occur at densities of approximately 2-5 deer per square mile. In contrast in Wisconsin, deer are found at 50+ animals per square mile. No one knows how rapidly CWD will spread among white-tailed deer at these densities or what long term affect this disease will have on a herd of this size (approximately 1.5 million animals). Chronic wasting disease is both transmissible and infectious, but specific details regarding transmission remain unknown. In contrast to BSE, CWD is not exclusively a foodborne disease associated with rendered ruminant meat, bonemeal, or animal protein products. Data from CWD epidemics in captive cervids and field data from wild cervids provide strong evidence that lateral transmission is the primary form of infection in susceptible animals. Vertical transmission, if it occurs, is likely relatively rare. Interspecific transmission probably occurs among the three susceptible native cervid species: mule deer, white-tailed deer, and elk. Rates of disease transmission are not well known in either captive or wild cervids. Based on modeling of field

16

16

data, an estimated 1.2-1.3 new infections per infectious animal per year occur for mule deer populations. These estimates were from relative low-density deer populations in Colorado and assume random, density-independent transmission among individuals. However, they indicated that transmission would be much higher (about 3.5 new infections per infectious animal per year) to simulate epidemics in captive deer populations, suggesting more intensive transmission under confinement or in high-density populations. No studies have been conducted to directly measure CWD transmission among wild cervids or to understand how social behavior and interaction among animals may affect disease transmission. Based on tissues from animals collected by hunters in the CWD-endemic area in Colorado, prevalence was similar between male and female mule deer (5.5% vs. 3.6%), white-tailed deer (2.3% vs. 1.4%), and elk (0.7% vs. 0.3%). For mule deer, prevalence differed between age classes within each sex. Prevalence was consistent across age class for females, but increased through the 4-6 year class then declined in males. In Wisconsin, prevalence of CWD in white-tailed deer was estimated at approximately 3% in the outbreak area, but prevalence at the center of the outbreak was estimated at approximately 13%. Little is known about the distribution of prevalence in white-tailed deer in relation to age, sex, or clinical disease status. Project Findings: Detailed project findings are presented below in the format of scientific papers. The paper on age, sex, and harvest patterns “Demographic Patterns and Harvest Vulnerability of Chronic Wasting Disease Infected White-tailed Deer in Wisconsin” has been published in the Journal of Wildlife Management. The paper on CWD transmission in adult females “Influence of genetic relatedness and spatial proximity on chronic wasting disease infection among female white-tailed deer” has been submitted for publication.

17

17

Research  Article  Demographic  Patterns  and  Harvest  Vulnerability  of  Chronic  Wasting  Disease  Infected  White-­‐Tailed  Deer  in  Wisconsin  DANIEL  A.  GREAR,1  Department  of  Wildlife  Ecology,  University  of  Wisconsin,  Madison,  WI  53706,  USA  MICHAEL  D.  SAMUEL,  U.S.  Geological  Survey—Wisconsin  Cooperative  Wildlife  Research  Unit,  University  of  Wisconsin,  Madison,  WI  53706,  USA  JULIE  A.  LANGENBERG,  Wisconsin  Department  of  Natural  Resources,  Madison,  WI  53707,  USA  DELWYN  KEANE,  Wisconsin  Veterinary  Diagnostic  Laboratory,  Madison,  WI  53705,  USA  Abstract  Chronic  wasting  disease  (CWD)  is  a  fatal  disease  of  white-­‐tailed  deer  (Odocoileus  virginianus)  caused  by  transmissible  protease-­‐resistant  prions.  Since  the  discovery  of  CWD  in  southern  Wisconsin  in  2001,  more  than  20,000  deer  have  been  removed  from  a  .2,500-­‐km2  disease  eradication  zone  surrounding  the  three  initial  cases.  Nearly  all  deer  removed  were  tested  for  CWD  infection  and  sex,  age,  and  harvest  location  were  recorded.  Our  analysis  used  data  from  a  310-­‐km2  core  study  area  where  disease  prevalence  was  higher  than  surrounding  areas.  We  found  no  difference  in  harvest  rates  between  CWD  infected  and  noninfected  deer.  Our  results  show  that  the  probability  of  infection  increased  with  age  and  that  adult  males  were  more  likely  to  be  infected  than  adult  females.  Six  fawns  tested  positive  for  CWD,  five  fawns  from  the  core  study  area,  including  the  youngest  (5  months)  free-­‐ranging  cervid  to  test  positive.  The  increase  in  male  prevalence  with  age  is  nearly  twice  the  increase  found  in  females.  We  concluded  that  CWD  is  not  randomly  distributed  among  deer  and  that  differential  transmission  among  sex  and  age  classes  is  likely  driving  the  observed  patterns  in  disease  prevalence.  We  discuss  alternative  hypotheses  for  CWD  transmission  and  spread  and,  in  addition,  discuss  several  possible  nonlinear  relationships  between  prevalence  and  age.  Understanding  CWD  transmission  in  free-­‐ranging  cervid  populations  will  be  essential  to  the  development  of  strategies  to  manage  this  disease  in  areas  where  CWD  is  found,  as  well  as  for  surveillance  strategies  in  areas  where  CWD  threatens  to  spread.  (JOURNAL  OF  WILDLIFE  MANAGEMENT  70(2):546–553;  2006)  

Key  words  Chronic  wasting  disease  (CWD),  disease  prevalence,  epidemiology,  harvest  vulnerability,  Odocoileus  virginianus,  prion,  transmissible  spongiform  encephalopathy  (TSE),  white-­‐tailed  deer,  Wisconsin.  The  discovery  of  chronic  wasting  disease  (CWD)  in  high-­‐density  white-­‐tailed  deer  (Odocoileus  virginianus)  populations  in  the  midwestern  and  eastern  United  States  has  increased  the  interest  of  wildlife  managers  in  understanding  the  epizootiology  of  this  disease,  its  potential  long-­‐term  impacts  on  deer  populations,  and  development  of  potential  management  strategies.  The  spread  of  CWD  threatens  areas  where  deer  hunting  is  an  important  cultural  and  economic  institution  and  is  an  essential  management  tool  for  controlling  high-­‐density  deer  populations.  Chronic  wasting  disease  (Williams  and  Young  1980)  belongs  to  a  family  of  diseases  known  as  transmissible  spongiform  encephalopathies  (TSEs),  which  affect  both  animals  (sheep  scrapie,  bovine  spongiform  encephalopathy,  transmissible  mink  encephalopathy)  and  humans  (Creutzfeldt–Jakob  disease  and  kuru).  The  causative  agent  of  TSEs  is  likely  an  abnormal  prion  protein  that  is  consistently  associated  with  the  disease  (Prusiner  1991).  Chronic  wasting  disease  is  the  only  infectious  TSE  that  affects  free-­‐ranging  cervid  species  including  elk  (Cervus  elaphus),  mule  deer  (Odocoileus  hemionus),  and  white-­‐tailed  deer  (Miller  et  al.  2000).  The  disease  was  first  recognized  in  captive  cervids  in  the  1960s,  and  since  1981  in  free-­‐ranging  cervids,  but  the  actual  length  of  time  that  the  condition  has  been  present  in  North  American  cervids  is  unknown.  Distribution  of  the  disease  in  North  America  is  largely  unknown  because  adequate  sampling  and  surveillance  have  not  been  conducted  in  most  areas  of  the  continent  (Samuel  et  al.  2003).  Until  2003,  CWD  was  found  in  free-­‐ranging  cervids  in  portions  of  Colorado,  Nebraska,  South  Dakota,  Wyoming,  Saskatchewan,  New  Mexico,  Illinois,  Utah,  and  Wisconsin  (Williams  and  Miller  2003).  Clinical  signs  develop  at  _1.5  years  

18

18

after  infection  in  wild  mule  deer  (Williams  et  al.  2002)  and  include  changes  in  behavior,  excessive  salivation,  periods  of  somnolence,  and  loss  of  body  condition.  Microscopic  spongiform  lesions  and  detection  of  abnormal  prion  protein  in  the  brain  accompany  clinical  signs.  No  captive  or  wild  cervid  has  ever  recovered  once  clinical  signs  develop  (Williams  et  al.  2002).  Similarities  between  lesions  and  epidemiology,  as  well  as  observation  of  interspecies  transmission,  indicate  that  the  same  CWD  agent  infects  all  three  species  (Williams  et  al.  2002).  In  addition,  similar  patterns  of  prevalence  related  to  age  and  sex  have  been  demonstrated  for  other  cervid  species  affected  by  CWD  (Miller  and  Conner  2005).  Chronic  wasting  disease  was  first  detected  in  south–central  Wisconsin  during  the  2001  fall  hunting  season,  but  the  source  of  CWD  infection  in  Wisconsin  remains  unknown  (Bartelt  et  al.  2003;  Joly  et  al.  2003).  In  2002,  the  Wisconsin  Department  of  Natural  Resources  (WDNR)  established  goals  of  eradicating  CWD  by  dramatically  reducing  white-­‐tailed  deer  density  within  a  1,064-­‐km2  area  surrounding  the  initial  CWD  infections  and  of  decreasing  the  probability  of  CWD  spread  by  reducing  the  deer  density  from  an  estimated  15–20  deer/km2  to  _4  deer/km2  in  areas  around  this  disease-­‐eradication  zone.  As  a  key  component  of  this  control  program,  retropharyngeal  lymph  nodes  and  brain  tissue  (obex)  from  deer  harvested  in  the  CWD  management  areas  were  collected  to  identify  new  CWD  infections,  assess  the  1  E-­‐mail:  dagrear@wisc.edu  546  The  Journal  of  Wildlife  Management  _  70(2)  distribution  of  CWD  on  the  landscape,  and  provide  data  for  research  on  CWD  epidemiology.  The  potential  impacts  of,  and  management  strategies  for  CWD  control  in  cervid  populations  remain  controversial.  Population  models  suggest  that  CWD  could  have  a  substantial  long-­‐term  impact  on  affected  populations  (Gross  and  Miller  2001;  J.  Cary,  University  of  Wisconsin,  unpublished  data).  Although  CWD  causes  direct  mortality  of  cervids,  the  long-­‐term  population  effects  of  the  disease  are  unknown.  In  addition,  public  concerns  regarding  human  risk  of  contracting  CWD  also  decrease  the  perceived  value  of  wild  cervids  and  affect  hunter  participation  (Petchenik  2003),  which  increases  the  difficulty  of  managing  high-­‐density  cervid  populations.  Currently,  there  is  no  evidence  that  CWD  will  spontaneously  disappear  or  be  controlled  without  management  intervention  (Gross  and  Miller  2001,  Peterson  et  al.  2002).  In  contrast,  there  is  significant  potential  for  expansion  of  the  geographic  range  of  the  disease,  and  once  established,  the  disease  could  be  maintained  through  environmental  contamination  for  an  unknown  period  of  time  (Peterson  et  al.  2002,  Miller  et  al.  2004).  Current  management  strategies  to  reduce  prevalence  or  eradicate  CWD  by  reducing  cervid  densities  assume  that  CWD  transmission  is  density  dependent  (Schauber  and  Wolf  2003)  and  homogeneous  among  animals.  Typically,  these  strategies  involve  surveillance  to  determine  the  prevalence  and  distribution  of  disease  and  intensive  culling  of  animals  within  the  affected  area  (Nebraska  Game  and  Parks  Commission  2002;  Williams  et  al.  2002;  Bartelt  et  al.  2003).  Studies  on  CWD  transmission  in  captive  deer  and  elk  indicate  that  lateral  transmission  by  direct  contact  and  ingestion  of  

19

19

abnormal  prion  via  contaminated  body  fluid  is  a  likely  mechanism  for  infection  (Williams  and  Young  1992;  Miller  et  al.  1998,  2000;  Miller  and  Williams  2003).  Vertical  transmission  does  not  seem  to  be  a  major  factor  in  transmission  (Gross  and  Miller  2001,  Miller  and  Williams  2003).  Transmission  from  infected  deer  to  the  environment,  then  to  susceptible  deer,  is  also  suspected  to  be  a  factor  in  transmission  (Miller  et  al.  2004),  but  the  mechanisms  for  this  route  of  transmission  and  their  significance  in  free-­‐ranging  cervids  are  not  understood.  In  addition,  the  effect  of  prion  dose,  genetic  resistance,  and  prion  strains  on  transmission  and  disease  progression  is  uncertain.  However,  recent  studies  on  prion  genetics  in  Wisconsin  white-­‐tailed  deer  indicate  that  .90%  of  the  population  genotypes  are  susceptible  to  disease  (Johnson  et  al.  2003).  These  unknown  factors  associated  with  routes  of  CWD  transmission  and  risks  of  disease  infection  related  to  density,  demography,  movement,  and  deer  behavior  have  hampered  the  development  of  management  strategies  and  public  acceptance  of  population  reduction  programs  to  eradicate  the  disease.  Demographic  patterns  of  CWD  infection  in  wild  cervids  can  indicate  when  disease  is  transmitted  among  animals,  which  animals  are  likely  to  become  infected,  and  how  disease  spreads  across  the  landscape.  However,  the  demographic  patterns  of  CWD  prevalence  in  white-­‐tailed  deer  related  to  age  and  sex  have  not  been  determined.  Our  major  objective  was  to  characterize  age  and  sex  prevalence  patterns  and  evaluate  hypotheses  about  CWD  transmission  in  free-­‐ranging  white-­‐tailed  deer.  Specifically,  we  tested  the  hypothesis  that  mass  action  (or  random)  transmission  of  CWD  occurs  in  white-­‐tailed  deer  populations  such  that  all  individuals  have  the  same  probability  of  becoming  infected.  Understanding  CWD  prevalence  patterns  may  also  help  improve  surveillance  programs  (Samuel  et  al.  2003)  and  disease  management  by  identifying  animals  that  are  most  likely  to  be  infected.  In  addition,  prevalence  patterns  from  harvested  mule  deer  in  Colorado  indicated  a  harvest  bias,  where  CWD  infected  animals  were  more  likely  to  be  harvested  as  the  harvest  season  progressed  (Conner  et  al.  2000).  Our  secondary  objective  was  to  test  for  differential  harvest  susceptibility  of  infected  white-­‐tailed  deer  to  evaluate  whether  harvest  bias  affected  our  prevalence  estimates  and  to  determine  the  potential  impacts  of  harvest  on  CWD  prevalence.  Methods  Study  Area  During  2002,  WDNR  conducted  spring  and  summer  culls  of  approximately  500  deer  to  obtain  a  preliminary  assessment  of  the  distribution  of  CWD  infection.  Based  on  this  initial  surveillance,  the  WDNR  established  a  1,064-­‐km2  CWD  eradication  zone  encompassing  all  positive  animals.  In  2003,  this  eradication  zone  was  expanded  to  cover  2,507  km2  as  new  positive  animals  were  detected  outside  the  2002  area  (Fig.  1).  For  our  analyses,  we  selected  a  310-­‐km2  core  study  area  within  the  disease  eradication  zone  where  the  highest  disease  prevalence  was  observed  (Joly  et  al.  2003,  fig.  1).  The  landscape  in  this  high-­‐prevalence  area  is  characterized  by  rolling  hills  and  small  stream  valleys  with  a  mixture  of  dairy  farms  and  oak-­‐hickory  woodlots,  almost  exclusively  in  private  ownership.  Prior  to  CWD  management  

20

20

efforts  (posthunt  2001)  deer  density  in  our  core  study  area  was  estimated  at  13.5–15.5  deer/km2  (Rolley  2002).  As  a  result  of  Figure  1.  Location  of  chronic  wasting  disease  (CWD)  management  areas  and  core  study  area  in  southern  Wisconsin,  USA.  The  core  study  area  (310  km2)  is  the  area  where  disease  prevalence  was  greatest  (6.7%)  in  2002  (Joly  et  al.  2003).  The  Wisconsin  Department  of  Natural  Resources  set  a  goal  of  eradicating  CWD  by  reducing  white-­‐tailed  deer  density  in  the  disease  eradication  zone.  The  disease-­‐eradication  zone  increased  in  size  from  1,064  to  2,507  km2  because  new  CWD  positives  were  found  to  the  west  and  south,  outside  of  the  2002  area.  The  herd-­‐reduction  zone  was  established  for  intensive  disease  surveillance  and  to  reduce  deer  densities  to  prevent  CWD  from  spreading  outside  of  the  infected  areas.  Grear  et  al.  _  Demographic  Patterns  of  CWD  547  generally  temperate  climate  and  abundant  resources,  white-­‐tailed  deer  in  this  area  have  high  fecundity  and  exhibit  very  little  seasonal  movement  (Larson  et  al.  1978,  Ishmael  1984).  Data  Collection  Deer  were  removed  by  hunter  harvest  and  limited  culling  by  government  sharpshooters  from  the  core  study  area  during  April  2002–April  2003  and  July  2003–January  2004.  Hunters  were  required  to  register  every  harvested  deer.  At  registration,  age,  sex,  and  kill  location  to  the  quarter  section  (0.65  km2)  were  recorded  for  each  deer.  Age  was  determined  by  tooth  replacement  and  wear  (Severinghaus  1949).  Heads  were  removed  from  harvested  deer  and  sent  to  a  tissue  extraction  center  where  a  portion  of  the  brain  stem  (obex)  and  retropharyngeal  lymph  nodes  (RPLN)  were  collected  for  CWD  diagnosis.  Tissues  from  all  deer  harvested  within  the  core  study  area  were  collected  from  the  2002  harvest,  but  primarily  adult  tissues  (.1  year  old)  were  collected  from  the  2003  harvest.  At  dissection,  a  portion  of  the  obex  and  1  RPLN  were  immediately  fixed  in  10%  buffered  formalin  and  the  remaining  tissues  were  frozen.  Retropharyngeal  lymph  nodes  and  obex  tissues  were  sent  to  the  Wisconsin  Veterinary  Diagnostic  Lab  for  CWD  testing  by  immunohistochemistry  (IHC)  (Miller  and  Williams  2002)  or  plate  ELISA  (IDEXX  Laboratories  Inc,  Westbrook,  Me.;  D.  Keane,  Wisconsin  Veterinary  Diagnostic  Laboratory,  personal  communication).  Fixed  RPLN  tissues  for  all  deer  from  the  2002  harvest  were  tested  with  the  use  of  IHC.  If  the  RPLN  tested  positive,  the  fixed  obex  tissue  was  also  tested  with  the  use  of  IHC.  Frozen  RPLN  tissues  from  a  majority  of  deer  harvested  during  2003  were  screened  with  the  IDEXX  test,  and  suspect  positives  were  confirmed  with  the  use  of  IHC  by  testing  RPLN  and  obex.  A  small  number  of  deer  harvested  in  2003  had  only  RPLN  and  obex  tissue  tested  by  IHC  (D.  Keane,  Wisconsin  Veterinary  Diagnostic  Laboratory,  personal  communication).  For  disease  reporting  and  demographic  analysis,  tests  showing  positive  IHC  reactions  in  the  RPLN  or  the  obex  were  classified  as  positive  for  CWD.  Statistical  Analysis  We  assessed  harvest  vulnerability  of  CWD  infected  deer  by  dividing  the  year  into  5  periods  that  roughly  corresponded  to  different  harvest  methods.  Period  1  (1  Apr–15  May)  corresponded  to  the  initial  surveillance,  Period  2  (16  May–23  Oct)  corresponded  to  summer  culls  targeted  at  high-­‐prevalence  areas  and  early  archery  season,  Period  3  (24  Oct–15  Nov)  corresponded  to  earlyseason  gun  harvest,  Period  4  (16–30  Nov)  corresponded  to  the  traditional  gun  harvest,  and  Period  5  (1  Dec–31  Mar)  corresponded  

21

21

to  extended  gun  harvest  and  sharpshooter  culls.  There  was  no  spring  surveillance  in  2003.  We  used  v2  analysis  (function  chisq,  R-­‐project  2004)  to  test  for  homogeneity  in  prevalence  between  periods  within  the  harvest  season,  and  we  used  a  Mantel–  Haenszel  test  for  trends  in  prevalence  during  the  harvest  season  (Schlessman  1982;  Freeman  1987).  We  performed  separate  analyses  for  all  deer,  as  well  as  on  bucks  and  does.  Conner  et  al.  (2000)  hypothesized  that  behavioral  change  due  to  CWD  infection  may  affect  vulnerability  to  harvest.  Therefore,  for  the  harvest-­‐vulnerability  analysis,  we  considered  individuals  to  be  positive  only  if  the  obex  was  infected  because  behavioral  changes  have  not  been  observed  prior  to  this  stage  of  disease.  However,  for  analysis  of  disease-­‐prevalence  patterns  based  on  age  and  sex,  we  considered  a  deer  positive  if  either  the  lymph  node  or  obex  was  infected.  We  used  a  Cochran–Mantel–Haenszel  v2  test  to  evaluate  yearto-­‐  year  change  in  prevalence  stratified  by  age  class  (function  mantelhaen.test,  R-­‐project  2004)  and  found  no  difference  in  prevalence  between  harvest  years.  As  a  result,  we  combined  prevalence  data  across  years  for  the  remainder  of  the  analyses.  We  examined  sex,  age,  and  year  as  possible  factors  that  affected  CWD  prevalence  with  the  use  of  logistic  regression  analysis  (function  glm,  R-­‐project  2004).  The  logistic  regression  model  for  predicting  prevalence  is  y  ¼  expðlÞ=ð1  þ  expðlÞÞ  where  y  is  the  prevalence  and  l  ¼  b0  þ  b1x1  þ  b2x2  þ  .  .  .  þ  bkxk  is  the  usual  linear  regression  that  includes  the  factors  that  affect  prevalence.  A  preliminary  analysis  indicated  an  interaction  between  the  affect  of  sex  and  age  on  prevalence,  so  we  conducted  separate  analyses  for  each  sex.  For  each  sex,  we  used  age  and  ln(age)  to  evaluate  potential  linear  and  nonlinear  trends  in  prevalence  with  age.  We  used  a  Hosmer  and  Lemeshow  goodness-­‐of-­‐fit  test  (Cox  and  Snell  1989)  to  assess  whether  our  models  were  a  good  fit  to  the  data  and  compared  alternative  logistic  regression  models  with  the  use  of  Akaike’s  Information  Criteria  (AIC;  Burnham  and  Anderson  1998).  To  attempt  clarification  of  the  nonlinear  patterns  with  age,  we  also  tested  for  a  decline  in  prevalence  in  older  age  classes  by  using  a  v2  test  that  compared  the  prevalence  in  peak  age  classes  to  all  older  deer.  To  compare  differences  in  prevalence  between  years,  ages,  and  sexes  we  calculated  the  odds  ratio  (OR)  and  95%  confidence  interval  (Selvin  1991)  of  a  deer  testing  positive  for  CWD.  Results  Harvest  Summary  Between  April  2002  and  January  2004,  21,285  deer  were  sampled  from  the  disease-­‐eradication  zone  and  316  (1.5%)  tested  positive.  In  our  core  study  area,  4,510  deer  were  sampled:  2,967  adults,  1,346  fawns,  and  197  deer  of  unknown  age  or  sex.  In  2002,  3,171  were  sampled  from  the  core  study  area  consisting  of  1,978  adults,  1,021  fawns,  and  172  deer  of  unknown  age  or  sex.  In  2003,  1,339  deer  were  sampled,  including  989  adults,  325  fawns,  and  25  deer  of  unknown  age  or  sex.  Estimated  prevalence  was  6.3%  (95%  CI:  5.5–7.2%)  for  adults  and  0.5%  (95%  CI:  0.1–0.9%)  for  fawns.  Estimated  adult  prevalence  was  6.7%  in  2002  and  5.3%  in  2003.  

22

22

Testing  results  were  also  categorized  by  RPLN  only  or  obex  and  RPLN  positive  for  each  individual.  Harvest-­‐vulnerability  analysis  was  performed  with  the  same  data  set  as  the  analysis  of  demographic  patterns,  but  this  analysis  considered  only  individuals  with  an  obex  positive  test  as  infected.  In  2002,  18  of  139  positive  tests  were  RPLN  positive  only,  including  15  of  134  positive  adults  and  3  of  5  positive  fawns  testing  RPLN  only  positive.  In  2003,  14  of  53  positive  tests  were  RPLN  only  positive.  Harvest  Vulnerability  Sixty-­‐three  percent  of  deer  were  harvested  during  the  early  and  traditional  hunting  seasons  in  October  and  November.  A  small  548  The  Journal  of  Wildlife  Management  _  70(2)  fraction  of  the  total  harvest  was  taken  during  spring  surveillance  in  2002,  about  15%  of  the  harvest  was  taken  during  summer,  and  15%  during  the  late  winter  periods  (Table  1).  For  harvest  vulnerability,  we  only  considered  animals  with  positive  CWD  test  in  their  obex  to  be  CWD  positive;  thus,  prevalence  shown  in  our  analysis  (Table  1)  is  slightly  lower  than  prevalence  based  on  lymph-­‐node  or  brain  infection.  We  found  no  evidence  that  CWD-­‐positive  deer  were  harvested  differentially  than  CWDnegative  deer.  Homogeneity  tests  indicated  that  adult  prevalence  did  not  vary  among  harvest  periods  during  2002  (v2¼8.40,  df¼4,  P  ¼  0.16)  and  2003  (v2  ¼  1.58,  df  ¼  3,  P  ¼  0.67),  for  adult  males  during  2002  (v2¼4.35,  df¼4,  P¼0.72)  and  2003  (v2¼0.11,  df¼  3,  P¼0.99),  or  for  adult  females  during  2002  (v2¼4.63,  df¼4,  P  ¼  0.40)  and  2003  (v2  ¼  5.21,  df  ¼  3,  P  ¼  0.53).  We  found  no  significant  difference  between  the  odds  of  infection  during  the  early  gun  season  (24  Oct–15  Nov)  and  traditional  gun  season  (16  Nov–30  Nov)  in  either  year  (2002:  OR  ¼1.616,  95%  CI  ¼0.70–  2.18,  v2  ¼  2.60,  2-­‐sided  P  ¼  0.21  df  ¼  1;  2003:  OR  ¼  0.91,  95%  CI  ¼  0.48–1.93,  v2  ¼  0.72,  2-­‐sided  P  ¼  0.42  df  ¼  1).  We  also  tested  for  a  linear  trend  in  prevalence  through  the  harvest  periods  using  a  Mantel–Haenzsel  test  but  found  no  apparent  trends  in  prevalence  for  adults  during  2002  (v2¼0.41,  df¼1,  P¼0.52)  and  2003  (v2¼0.11,  df¼1,  P¼0.74),  for  adult  males  during  2002  (v2  ¼0.26,  df¼1,  P¼0.61)  and  2003  (v2¼0.11,  df¼1,  P¼0.74),  or  for  adult  females  during  2002  (v2  ¼  0.09,  df  ¼  1,  P  ¼  0.76)  and  2003  (v2  ¼  0.36,  df  ¼  1,  P  ¼  0.55).  Demographic  Patterns  Five  of  1,021  (0.5%)  fawns  tested  positive  in  2002  and  zero  of  325  tested  positive  in  2003  in  the  core  study  area  (1  additional  fawn  in  the  surrounding  lower  prevalence  area  also  tested  positive).  Three  of  these  fawns,  approximately  5  months  old  (2  animals)  and  9  months  old,  tested  positive  in  RPLN  only.  The  remaining  2  fawns,  approximately  9  and  10  months  old,  tested  positive  in  RPLN  and  obex.  For  all  the  following  prevalence  data  and  analyses,  we  defined  positive  cases  as  having  CWD  infection  in  the  RPLN  or  obex.  Adult  prevalence  in  2002  (6.8%)  was  similar  to  2003  (5.4%;  OR¼1.26,  95%  CI  ¼0.91–1.75)  and  prevalence  in  fawns  was  lower  than  in  adults  (OR  ¼  0.0523,  95%  CI  ¼  0.022–0.127).  We  excluded  fawns  from  further  demographic  analysis  because  their  prevalence  was  so  much  lower  than  adult  deer.  Adult  males  had  a  higher  overall  prevalence  (7.4%)  than  females  (5.4%;  OR¼1.43,  95%  CI¼1.07–1.91;  Table  1).  Adult  male  prevalence  in  2002  (8.1%)  was  similar  to  2003  (6.4%)  (OR  

23

23

¼  1.29,  95%  CI  ¼  0.84–1.97).  Adult  female  prevalence  was  also  similar  between  2002  (5.8%)  and  2003  (4.2%;  OR¼  1.09,  95%  CI  ¼  0.65–1.82).  Prevalence  between  years  did  not  differ  when  stratified  by  age  in  males  (Cochran–Mantel–Haenszel  v2¼0.734,  df  ¼  1,  P  ¼  0.39)  or  females  (Cochran–Mantel–Haenszel  v2  ¼  1.10,  df  ¼  1,  P  ¼  0.29).  Sex-­‐specific  logistic  regression  models  of  prevalence  show  that  age  has  a  nonlinear  relationship  with  prevalence.  Our  regression  models  with  asymptotic  or  declining  trend  in  prevalence  at  older  ages  fit  better  than  regression  models  with  a  linear  trend  in  prevalence  with  age  (Fig.  2).  However,  we  found  little  support  (DAIC  ,  2.0)  for  distinguishing  between  an  asymptotic  and  declining  relationship  between  prevalence  and  older  age  classes  for  either  males  (Table  2)  or  females  (Table  3).  In  addition,  there  was  no  evidence  that  peak  prevalence  in  the  3-­‐year-­‐old  age  class  was  different  from  prevalence  in  older  age  classes  for  males  (v2  ¼  0.003,  df¼1,  P¼0.95)  or  females  (v2¼0.144,  df¼1,  P¼0.70).  Discussion  Demographic  Patterns  Our  results  showed  that  the  risk  of  CWD  infection  does  not  follow  a  random  transmission  process,  which  predicts  homogeneous  rates  of  infection  among  age  and  sex  classes.  We  found  a  significant  increase  in  prevalence  with  age  for  both  male  and  female  white-­‐tailed  deer.  This  pattern  is  characteristic  of  a  chronic  disease,  like  CWD,  where  the  cumulative  risk  of  infection  increases  with  length  of  potential  exposure  (age).  We  found  that  CWD  prevalence  was  3–4%  for  yearling  males  and  females,  but  increased  to  peak  prevalence  in  3-­‐year-­‐olds  of  both  genders.  However,  peak  prevalence  in  3-­‐year-­‐old  males  (13%)  was  nearly  twice  that  for  females  (7%),  illustrating  the  dramatic  increase  in  risk  of  infection  for  males  after  the  yearling  class.  We  believe  these  infection  patterns  reflect  heterogeneous  rates  and/or  pathways  of  disease  transmission  in  Wisconsin  based  on  the  social  behavior  of  white-­‐tailed  deer.  Miller  and  Conner  (2005)  observed  similar  age  and  sex  patterns  in  a  separate  CWD  epidemic  in  Colorado  mule  deer,  and  O’Brien  et  al.  (2002)  observed  similar  epidemiology  in  Michigan  in  white-­‐tailed  deer  infected  with  bovine  tuberculosis  (TB;  a  chronic  bacterial  infection).  Both  studies  hypothesized  that  social  behavior  is  a  strong  explanation  for  these  infection  patterns.  Table  1.  Number  of  deer  harvested  and  prevalence  of  obex-­‐positive  individuals  by  sex  in  5  periods  during  2002  and  2003  from  the  chronic  wasting  disease  (CWD)  high-­‐prevalence  area  in  southern  Wisconsin,  USA.  Initial  CWD  surveillance  was  conducted  in  spring  2002  but  was  not  repeated  in  spring  2003.  Period  2002  Males  Females  2003  Males  Females  n  Prevalence  n  Prevalence  n  Prevalence  n  Prevalence  n  Prevalence  n  Prevalence  Spring  (1  Mar–15  May)  193  0.052  39  0.051  154  0.052  0  0  0  0  0  0  Summer  (1  Jun–23  Oct)  388  0.069  124  0.081  264  0.064  59  0.068  23  0.043  36  0.083  Early  fall  (24  Oct–15  Nov)  679  0.059  385  0.068  294  0.048  563  0.036  353  0.045  210  0.019  Late  fall  (15  Nov–30  Nov)  395  0.035  166  0.042  229  0.031  228  0.044  101  0.050  127  0.040  Winter  (1  Dec–28  Feb)  323  0.087  114  0.105  209  0.077  139  0.036  55  0.055  84  0.024  Total  1978  0.060  828  0.069  1150  0.054  989  0.039  532  0.047  457  0.031  Grear  et  al.  _  Demographic  Patterns  of  CWD  549  

24

24

We  believe  that  different  social  structure  and  behavior  of  bucks  and  does  may  be  important  factors  causing  differential  CWD  infection.  Females  form  matrilineal  social  units  comprised  of  an  older  female,  her  daughters,  and  up  to  4  generations  of  closely  related  females  (Hawkins  and  Klimstra  1970).  These  groups  are  highly  philopatric  to  summer  and  winter  ranges  and  to  related  members  of  the  group  (Mathews  and  Porter  1993,  Aycrigg  and  Porter  1997,  Nelson  and  Mech  1999).  In  southern  Wisconsin,  females  use  relatively  small  home  ranges  (1.5–2.5  km2)  and  move  little  throughout  the  year  (Larson  et  al.  1978,  Ishmael  1984).  In  contrast  to  does,  bucks  have  larger  home  ranges  (2–4  km2),  form  smaller  male  social  units  that  are  seasonally  dynamic,  seldom  philopatric,  and  have  little  contact  with  females  for  most  of  the  year  (Hirth  1976,  Wozencraft  1978,  Nixon  et  al.  1991).  The  composition  of  buck  groups  varies  from  larger  groups  in  late  winter  and  early  spring  to  individual  males  with  increased  movement  and  contact  with  competing  bucks  and  females  during  the  breeding  season  (Hirth  1976,  Nixon  1991).  Based  on  differences  in  male  social  behavior  and  movement,  we  suggest  several  hypotheses  to  explain  the  increased  risk  of  CWD  infection  in  males  compared  to  females.  First,  direct  transmission  may  occur  among  males  in  buck  groups  from  late  winter  through  summer.  Second,  transmission  may  occur  during  the  breeding  season  when  susceptible  bucks  contact  many  infected  females  or  when  bucks  visit  scent  stations  (rubs  and  scrapes)  used  by  infected  bucks.  Third,  bucks  have  a  greater  chance  to  contact  CWD  in  the  environment  than  susceptible  females  due  to  their  larger  home  range  size  and  breeding  season  movements.  Fourth,  females  may  be  more  susceptible  to  disease  mortality  and/or  males  may  have  a  longer  preclinical  period  than  females.  Behavioral  evidence  supports  the  rationale  for  the  first  3  hypotheses  (Hirth  1976,  Nixon  1991),  whereas  there  is  no  evidence  to  suggest  that  males  or  females  have  different  susceptibility  to  CWD.  Prevalence  of  CWD  in  adult  white-­‐tailed  deer  appears  to  approach  an  asymptote  or  potentially  decline  in  older  age  classes  (Fig.  2),  but  we  were  unable  to  distinguish  between  these  alternative  patterns.  We  believe  the  trends  in  prevalence  for  older  deer  provide  important  epidemiological  information  about  the  disease  progression  and  the  potential  effects  of  CWD  on  cervid  populations.  For  chronic  diseases,  age-­‐specific  prevalence  will  generally  be  determined  by  the  infection  rate  and  the  disease  mortality  rate.  Increasing  prevalence  typically  occurs  with  age  because  infection  rate  exceeds  disease  mortality  rate.  But  for  older  individuals,  the  number  dying  from  the  disease  can  reach  equilibrium  with  the  number  of  new  infections,  resulting  in  a  constant  prevalence  in  the  older  age  classes.  If  mortality  rates  exceed  infection  rates,  the  result  would  be  decreasing  prevalence  in  older  age  classes.  Declining  prevalence  in  older  age  classes  may  indicate  that  infection  rates  are  decreasing  or  disease  mortality  rates  are  increasing.  Studies  on  prion  diseases  in  sheep  (Redman  et  al.  2002),  humans  (Bacchetti  2003),  and  mule  deer  (Miller  et  al.  2000,  Miller  and  Figure  2.  Observed  chronic  wasting  disease  (CWD)  prevalence  and  estimated  prevalence  from  alternative  logistic  regression  models  based  on  age  in  male  and  female  adult  white-­‐tailed  deer  from  southern  Wisconsin,  USA,  Apr  2002–  Feb  2004.  

25

25

Table  2.  Alternative  logistic  regression  models  of  chronic  wasting  disease  (CWD)  prevalence  in  adult  male  white-­‐tailed  deer  (n  ¼  1360)  from  southern  Wisconsin,  USA,  Mar  2002–Jan  2004.  Model  df  v2  a  pb  DAIC  xi  c  

Age  þ  ln(age)  2  0.068  0.99  0  0.58  ln(age)  3  1.417  0.23  0.74  0.4  Age  3  5.334  0.15  6.95  0.03  a  v2  is  the  Hosmer  and  Lemeshow  ratio  goodness-­‐of-­‐fit  test  statistic.  b  p  is  the  significance  of  the  goodness-­‐of-­‐fit  test.  c  xi  refers  to  the  relative  probability  that  the  model  is  the  best  model  given  the  data.  550  The  Journal  of  Wildlife  Management  _  70(2)  Conner  2005)  have  shown  that  prevalence  peaks  and  then  declines  with  age.  Heisey  and  Joly  (2004)  proposed  that  one  explanation  for  this  pattern  for  prion  diseases  could  be  decreased  immune  function  in  older  individuals.  They  postulated  that  healthy  immune  function  may  facilitate  the  progression  of  infectious  prions  through  lymphatic  tissues,  whereas  immune  system  senescence  may  provide  resistance.  This  hypothesis  predicts  that  decreased  infection  rates  in  older  individuals  would  produce  lower  observed  prevalence.  Alternatively,  if  disease  mortality  rates  increased  in  older  age  classes,  prevalence  would  also  decline.  Although  this  is  a  typical  pattern  for  a  conventional  infectious  disease  due  to  reduced  immunity  in  older  individuals,  CWD  does  not  initiate  an  immune  response  like  a  typical  infectious  agent  (Williams  et  al.  2002).  We  suggest  that  spatial  heterogeneity  of  CWD  infection  (or  temporal  heterogeneity  of  sample  collection)  also  could  produce  declining  prevalence  patterns  in  older  animals  because  disease  mortality  in  infected  landscape  patches  may  result  in  a  declining  age  structure,  whereas  an  older  age  structure  could  persist  in  uninfected  patches.  Because  the  white-­‐tailed  deer  population  in  southern  Wisconsin  has  a  relatively  young  age  structure,  it  is  difficult  for  us  to  determine  exactly  what  pattern  may  be  occurring  in  older  age  classes.  More  research  is  necessary  to  determine  whether  heterogeneity,  infection,  or  mortality  rates  are  involved  in  these  patterns.  Harvest  Vulnerability  Conner  et  al.  (2000)  reported  that  CWD  prevalence  in  hunterharvested  mule  deer  in  Colorado  increased  during  the  hunting  season.  They  considered  seasonal  movement  of  deer  from  areas  with  less  hunter  harvest  to  areas  with  more  harvest  as  a  likely  cause  for  this  bias.  They  also  concluded  that  differential  harvest  vulnerability  of  infected  animals  due  to  behavioral  changes  caused  by  disease  was  a  less  likely  cause  for  the  bias.  In  contrast,  we  detected  no  seasonal  trends  in  our  analyses  of  harvested  Wisconsin  white-­‐tailed  deer.  Because  the  deer  herd  in  southern  Wisconsin  showed  little  or  no  seasonal  movement,  our  results  were  not  confounded  by  seasonal  migration  patterns  like  those  found  in  western  mule  deer.  In  addition,  Spraker  et  al.  (1997)  reported  there  was  little  evidence  that  disease  progression  or  behavioral  changes  due  to  CWD  infection  were  different  for  mule  deer  than  white-­‐tailed  deer.  Thus,  it  seems  unlikely  that  CWD  infection  causes  substantial  differential  susceptibility  of  white-­‐tailed  or  mule  deer  to  hunter  harvest.  These  results  indicate  that  harvest  can  be  used  as  effective  tool  for  collecting  cervids  for  estimating  CWD  prevalence  rates,  transmission  rates,  and  other  epidemiological  

26

26

parameters.  Fawns  and  Yearlings  Infection  in  Wisconsin  white-­‐tailed  deer  fawns  between  5  and  10  months  of  age  represented  the  youngest  free-­‐ranging  cervids  to  test  positive  for  CWD.  In  disease-­‐progression  studies  CWD  can  be  detected  in  RPLN  tissue  of  captive  mule  deer  fawns  by  2  months  postinoculation  (Sigurdson  et  al.  1999).  These  results  suggest  that  the  5-­‐month-­‐old  deer  were  likely  infected  prior  to  weaning.  This  could  indicate  transmission  in  utero  or  transmission  within  several  months  of  birth.  However,  pen  studies  on  CWD  indicate  that  transmission  in  utero  is  not  a  likely  route  of  infection  (Williams  and  Miller  2003).  White-­‐tailed  deer  fawns  use  very  small  ranges  and  engage  in  intense  grooming  and  nursing  exclusively  with  their  mother  for  the  first  6  weeks  postparturition  (Schwede  et  al.  1993).  Therefore,  it  is  likely  that  transmission  within  2–3  months  following  birth  is  from  contact  with  the  mother  or  from  contact  with  CWD-­‐contaminated  surroundings  in  the  mother’s  home  range.  Because  CWD  prevalence  in  fawns  (0.5%)  was  low  compared  to  the  prevalence  in  adult  does  (5.4%),  our  data  indicated  a  low  probability  of  CWD  being  transmitted  from  infected  does  to  their  fawns  early  in  life.  If  mother-­‐to-­‐offspring  transmission  was  common,  especially  prior  to  weaning,  we  would  expect  higher  prevalence  in  fawns.  Alternatively,  if  we  failed  to  detect  early  stages  of  infection  in  some  fawns,  we  would  expect  increased  infection  in  yearlings  following  an  additional  year  of  disease  progression  and  exposure.  However,  yearling  prevalence  was  also  much  lower  than  adult  female  prevalence.  Our  results  for  freeranging  white-­‐tailed  deer  support  research  by  Miller  and  Williams  (2003)  in  captive  mule  deer  that  maternal  (dam  to  young)  transmission  is  not  an  important  route  for  CWD  infection.  Based  on  their  low  prevalence  rates,  fawns  are  unlikely  to  play  an  important  role  in  maintaining  and  spreading  CWD.  Difference  in  movement  and  dispersal  between  white-­‐tailed  bucks  and  does  may  also  be  a  significant  component  of  CWD  distribution  across  the  landscape,  especially  in  areas  where  animals  do  not  show  seasonal  migration.  Yearling  male  dispersal  rates  are  50%  to  .80%  with  dispersal  distances  between  10  and  30  km,  compared  with  female  dispersal  rates  of  ,20%  (Hawkins  and  Klimstra  1970,  Nixon  et  al.  1991,  Nelson  1993,  Rosenberry  et  al.  1999).  Infected  yearling  males  have  the  potential  to  spread  the  disease  over  a  large  geographic  area.  Fortunately,  prevalence  in  yearling  males  and  females  is  similar,  3.4%  and  3.3%,  respectively,  and  considerably  lower  than  adult  males.  Management  Implications  An  understanding  of  processes  that  drive  transmission  of  CWD  among  free-­‐ranging  deer  and  cause  the  geographic  spread  of  disease  will  aid  in  developing  effective  strategies  for  CWD  management  in  white-­‐tailed  deer  and  other  cervid  populations.  Our  results  provide  biologically  based  hypotheses  about  the  mechanisms  of  CWD  transmission,  but  do  not  provide  sufficient  information  to  distinguish  the  relative  importance  of  direct  CWD  transmission  by  animal-­‐to-­‐animal  contact  and  indirect  (environmental)  routes  of  transmission.  We  recommend  further  research  Table  3.  Alternative  logistic  regression  models  of  chronic  wasting  disease  

27

27

(CWD)  prevalence  in  adult  female  white-­‐tailed  deer  (n  ¼  1607)  from  southern  Wisconsin,  USA,  Mar  2002–Jan  2004.  Model  df  v2  a  pb  DAIC  xi  c  

Age  þ  ln(age)  2  0.072  0.99  0  0.51  ln(age)  3  2.621  0.45  0.63  0.37  Age  3  4.861  0.18  2.9  0.12  a  v2  is  the  Hosmer  and  Lemeshow  goodness-­‐of-­‐fit  test  statistic.  b  p  is  the  significance  of  the  goodness-­‐of-­‐fit  test.  c  xi  refers  to  the  relative  probability  that  the  model  is  the  best  model  given  the  data.  Grear  et  al.  _  Demographic  Patterns  of  CWD  551  regarding  how  contact  among  animals  or  accumulation  of  environmental  sources  of  infectious  prions  influence  transmission.  In  addition,  we  note  that  most  challenge  studies  in  cervids  have  been  conducted  with  young  animals,  and  we  suggest  that  similar  studies  using  older  animals  would  provide  a  useful  comparison  given  the  apparent  higher  risk  of  infection  in  older  deer  in  the  wild.  Given  the  low  observed  prevalence  in  fawns,  we  recommend  that  surveillance  programs  with  the  goal  of  detecting  disease  where  it  has  not  been  previously  found  should  focus  testing  on  animals  .12–18  months  of  age.  Surveillance,  with  the  goal  of  detecting  geographic  spread  of  CWD  from  an  infected  area,  should  focus  on  yearling  and  adult  bucks,  as  these  animals  are  the  most  likely  to  disperse  the  disease  by  natural  means.  Finally,  we  recommend  that  in  areas  where  disease  is  present  and  the  goal  is  to  estimate  prevalence,  testing  should  focus  on  adult  deer  of  both  sexes.  We  suggest  a  strategy  that  focuses  on  removing  CWD-­‐positive  animals,  along  with  density  reduction,  in  situations  where  controlling  CWD  is  the  goal.  Adult  bucks  with  larger  home  ranges  are  much  more  likely  to  be  infected  than  either  young  dispersing  bucks  or  adult  does,  and  thus  create  the  greatest  potential  for  local  disease  spread  as  well  as  the  highest  probability  for  removal  of  infected  animals.  Although  does  have  a  lower  disease  prevalence,  they  are  likely  to  be  more  abundant  than  adult  bucks,  due  to  higher  hunting  pressure  on  large  bucks,  hunting  traditions,  and  regulations  that  tend  to  protect  does.  As  a  result  of  these  competing  implications  and  in  the  absence  of  clear  information  about  the  routes  of  CWD  transmission,  management  strategies  need  to  strike  a  balance  among  efforts  to  reduce  deer  population  density  by  increasing  antlerless  harvests,  reducing  prevalence  and  density  of  infected  animals  in  highly  affected  areas  by  removing  older  females  and  males,  and  eliminating  the  spread  of  disease  to  new  areas  by  removing  males  of  all  ages.  In  the  long  term,  further  research,  preferably  integrated  with  management  actions,  will  be  needed  to  develop  a  better  knowledge  of  the  factors  affecting  the  transmission  of  CWD  in  free-­‐ranging  cervids  and  to  develop  appropriate  management  strategies.  Acknowledgments  This  study  could  not  have  been  completed  without  the  time  and  hard  work  of  the  Wisconsin  Department  of  Natural  Resources  staff  and  volunteers  and  the  Wisconsin  Veterinary  Diagnostic  Lab  TSE  testing  staff.  We  also  thank  C.  Batha,  T.  Howard,  M.  Watrud,  K.  Beheler,  and  J.  Sausen  of  the  WDNR  and  P.  Boschler  of  the  Wisconsin  Veterinary  Diagnostic  Laboratory  for  the  time  and  effort  they  put  into  sample  collection,  data  management,  and  

28

28

disease  testing.  Funding  was  provided  through  the  United  States  Geological  Survey  National  Wildlife  Health  Center  and  the  Wisconsin  Department  of  Natural  Resources.  We  also  acknowledge  D.  Joly,  R.  Rolley,  and  M.  Miller  for  valuable  comments.  Literature  Cited  Aycrigg,  J.  L.,  and  W.  F.  Porter.  1997.  Sociospatial  dynamics  of  white-­‐tailed  deer  in  the  central  Adirondack  Mountains,  New  York.  Journal  of  Mammalogy  78:468–482.  Bacchetti,  P.  2003.  Age  and  variant  Creuztfeldt–Jakob  disease.  Emerging  Infectious  Diseases  9:1611–1612.  Bartelt,  G.,  J.  Pardee,  and  K.  Thiede.  2003.  Environmental  impact  statement  on  rules  to  eradicate  chronic  wasting  disease  in  Wisconsin’s  free-­‐ranging  white-­‐tailed  deer  herd.  Wisconsin  Department  of  Natural  Resources,  Madison,  USA.  Burnham,  K.  P.,  and  D.  R.  Anderson.  1998.  Model  selection  and  inference:  a  practical  information-­‐theoretic  approach.  Spring-­‐Verlag,  New  York,  New  York,  USA.  Conner,  M.  M.,  C.  W.  McCarty,  and  M.  W.  Miller.  2000.  Detection  of  bias  in  harvest-­‐based  estimates  of  chronic  wasting  disease  prevalence  in  mule  deer.  Journal  of  Wildlife  Diseases  36:691–699.  Cox,  D.  R.,  and  E.  J.  Snell.  1989.  Analysis  of  binary  data.  Second  edition.  Chapman  and  Hall,  New  York,  New  York,  USA.  Freeman,  D.  H.  1987.  Applied  categorical  data  analysis.  Marcel  Dekker,  New  York,  New  York,  USA.  Gross,  J.  E.,  and  M.  W.  Miller.  2001.  Chronic  wasting  disease  in  mule  deer:  disease  dynamics  and  control.  Journal  of  Wildlife  Management  65:205–215.  Hawkins,  R.  E.,  and  W.  D.  Klimstra.  1970.  A  preliminary  study  of  the  social  organization  of  white-­‐tailed  deer.  Journal  of  Wildlife  Management  34:407–  419.  Heisey,  D.  M.,  and  D.  O.  Joly.  2004.  Age  and  transmissible  spongiform  encephalopathies.  Emerging  Infectious  Diseases  10:1164–1165.  Hirth,  D.H.  1976.  Social  behavior  of  white-­‐tailed  deer  in  relation  to  habitat.  Wildlife  Monographs  53.  Ishmael,  W.  E.  1984.  White-­‐tailed  deer  ecology  and  management  in  Southern  Wisconsin.  Thesis,  University  of  Wisconsin,  Madison,  USA.  Johnson,  C.,  J.  Johnson,  M.  Clayton,  D.  McKenzie,  and  J.  Aiken.  2003.  Prion  protein  gene  heterogeneity  in  free-­‐ranging  white-­‐tailed  deer  within  the  chronic  wasting  disease  affected  region  of  Wisconsin.  Journal  of  Wildlife  Diseases  39:576–581.  Joly,  D.  O.,  C.  A.  Ribic,  J.  A.  Langenberg,  K.  Beheler,  C.  A.  Batha,  B.  J.  Dhuey,  R.  E.  Rolley,  G.  Bartelt,  T.  R.  Van  Deelen,  and  M.  D.  Samuel.  2003.  Chronic  wasting  disease  in  free-­‐ranging  Wisconsin  white-­‐tailed  deer.  Emerging  Infectious  Diseases  9:599–601.  Larson,  T.  J.,  O.  J.  Rongstad,  and  F.  W.  Tebilcox.  1978.  Movement  and  habitat  use  of  white-­‐tailed  deer  in  south  central  Wisconsin.  Journal  of  Wildlife  Management  42:113–117.  Mathews,  N.  E.,  and  W.  F.  Porter.  1993.  Effect  of  social  structure  on  genetic  structure  of  free-­‐ranging  white-­‐tailed  deer  in  the  Adirondack  Mountains.  Journal  of  Mammalogy  74:33–43.  Miller,  M.  W.,  M.  A.  Wild,  and  E.  S.  Williams.  1998.  Epidemiology  of  chronic  wasting  disease  in  Rocky  Mountain  elk.  Journal  of  Wildlife  Diseases  34:532–  538.  Miller,  M.  W.,  E.  S.  Williams,  C.  W.  McCarty,  T.  R.  Spraker,  T.  J.  Kreeger,  C.  T.  Larsen,  and  E.  T.  Thorne.  2000.  Epizootiology  of  chronic  wasting  disease  in  free-­‐ranging  cervids  in  Colorado  and  Wyoming.  Journal  of  Wildlife  Diseases  36:676–690.  Miller,  M.  W.,  and  E.  S.  Williams.  2002.  Detection  of  PrPcwd  in  mule  deer  by  immunohistochemistry  of  lymphoid  tissues.  Veterinary  Record:  610–612.  Miller,  M.  W.,  and  E.  S.  Williams.  2003.  Horizontal  prion  transfer  in  mule  deer.  Nature  425:35–36.  Miller,  M.  W.,  E.  S.  Williams,  N.  T.  Hobbs,  and  L.  L.  Wolfe.  2004.  Environmental  sources  of  prion  transmission  in  mule  deer.  Emerging  Infectious  Diseases  10:1003–1006.  Miller,  M.  W.,  and  M.  M.  Conner.  2005.  Epidemiology  of  chronic  wasting  disease  in  free-­‐ranging  mule  deer:  spatial,  temporal,  and  demographic  influences  on  observed  prevalence.  Journal  of  Wildlife  Diseases  41:275–290.  Nebraska  Game  and  Parks  Commission.  2002.  CWD  test  results,  northern  

29

29

Sioux  County,  NE.  Sampling  periods:  November  2001  and  January/February  2002.  http://www.ngpc.state.ne.us/wildlife/cwd/cwdFAQ.html.  Accessed  2002  Nov  12.  Nelson,  M.  E.  1993.  Natal  dispersal  and  gene  flow  in  white-­‐tailed  deer  in  northeastern  Minnesota.  Journal  of  Mammalogy  74:316–322.  Nelson,  M.  E.,  and  L.  D.  Mech.  1999.  Twenty-­‐year  home  range  dynamics  of  a  white-­‐tailed  deer  matriline.  Canadian  Journal  of  Zoology  77:1128–1135.  Nixon,  C.  M.,  L.  P.  Hansen,  P.  A.  Brewer,  and  J.  E.  Chelsvig.  1991.  Ecology  of  white-­‐tailed  deer  in  an  intensively  farmed  region  of  Illinois.  Wildlife  Monographs  No.  118.  O’Brien,  D.  J.,  S.  M.  Schmitt,  J.  S.  Fierke,  S.  A.  Hogle,  S.  R.  Winterstein,  T.  M.  Cooley,  W.  E.  Moritz,  K.  L.  Diegel,  S.  D.  Fitzgerald,  D.  E.  Berry,  and  J.  B.  Kaneene.  2002.  Epidemiology  of  Mycobacterium  bovis  in  free  ranging  white-­‐  552  The  Journal  of  Wildlife  Management  _  70(2)  tailed  deer,  Michigan,  USA,  1995–2000.  Preventive  Veterinary  Medicine  54:  47–63.  Petchenik,  J.  2003.  Chronic  wasting  disease  in  Wisconsin  and  the  2002  hunting  season:  gun  deer  hunters’  first  response.  Wisconsin  Department  of  Natural  Resources:  Bureau  of  Integrated  Science  Services.  Miscellaneous  publication  PUB-­‐SS-­‐982  2003.  Madison,  Wisconsin,  USA.  Peterson,  M.  J.,  M.  D.  Samuel,  V.  F.  Nettles,  Jr.,  G.  Wobeser,  and  W.  D.  Hueston.  2002.  Review  of  chronic  wasting  disease  management  policies  and  programs  in  Colorado.  Colorado  Wildlife  Commission,  Denver,  USA.  Prusiner,  S.  B.  1991.  Molecular  biology  of  prion  diseases.  Science  252:1515–  1522.  Redman,  C.  A.,  P.  G.  Coen,  L.  Matthews,  R.  M.  Lewis,  W.  S.  Dingwall,  J.  D.  Foster,  M.  E.  Chase-­‐Topping,  N.  Hunter,  and  M.  E.  J.  Woolhouse.  2002.  Comparative  epidemiology  of  scrapie  outbreaks  in  individual  sheep  flocks.  Epidemiology  and  Infection  128:513–521.  Rolley,  R.  2002.  White-­‐tailed  deer  population  status  2001.  In:  Wisconsin  Wildlife  Surveys.  Compiled  by  B.  Dhuey  and  H.  Arrowood.  Wisconsin  Department  of  Natural  Resources.  Bureau  of  Integrated  Science  Services.  Miscellaneous  Publication  PUB-­‐SS-­‐970  04/2002.  Madison,  Wisconsin,  USA.  Rosenberry,  C.  S.,  R.  A.  Lancia,  and  M.  C.  Conner.  1999.  Population  effects  of  white-­‐tailed  deer  dispersal.  Wildlife  Society  Bulletin  27:858–864.  [R-­‐project]  The  R  project  for  statistical  computing.  2004  Nov  28.  R-­‐project  home  page:  ,http://www.r-­‐project.org..  Accessed  2004  Nov  28.  Samuel,  M.  D.,  D.  O.  Joly,  M.  A.  Wild,  S.  D.  Wright,  D.  L.  Otis,  R.  W.  Werge,  and  M.  W.  Miller.  2003.  Surveillance  strategies  for  detecting  chronic  wasting  disease  in  free-­‐ranging  deer  and  elk.  USGS–National  Wildlife  Health  Center,  Madison,  Wisconsin.  Schauber,  E.  M.,  and  A.  Woolf.  2003.  Chronic  wasting  disease  in  deer  and  elk:  A  critique  of  current  models  and  their  application.  Wildlife  Society  Bulletin  31:  610–616.  Schlessman,  J.  J.  1982.  Case-­‐control  studies.  Oxford  University  Press,  New  York,  USA.  Schwede,  G.,  H.  Hendrichs,  and  W.  McShea.  1993.  Social  and  spatial  organization  of  female  white-­‐tailed  deer,  Odocoileus  virginianus,  during  the  fawning  season.  Animal  Behaviour  45:1007–1017.  Selvin,  S.  1991.  Statistical  analysis  of  epidemiological  data.  Oxford  University  Press,  New  York,  USA.  Severinghaus,  C.  W.  1949.  Tooth  development  and  wear  as  criteria  of  age  in  white-­‐tailed  deer.  Journal  of  Wildlife  Management  13:195–216.  Sigurdson,  C.  J.,  E.  S.  Williams,  M.  W.  Miller,  T.  I.  Spraker,  K.  L.  O’Rourke,  and  E.  A.  Hoover.  1999.  Oral  transmission  and  early  lymphoid  tropism  of  chronic  wasting  disease  PrPres  in  mule  deer  fawns  (Odocoileus  hemionus).  Journal  of  General  Virology  80:2757–2764.  Spraker,  T.  R.,  M.  W.  Miller,  E.  S.  Williams,  D.  M.  Getzy,  W.  J.  Adrian,  G.  G.  Schooonveld,  R.  A.  Spowart,  K.  I.  O’Rourke,  J.  M.  Miller,  and  P.  A.  Merz.  1997.  Spongiform  encephalopathy  in  free-­‐ranging  mule  deer  (Odocoileus  hemionus),  white-­‐tailed  deer  (Odocoileus  virginianus)  and  Rocky  Mountain  elk  (Cervus  elaphus  nelsoni)  in  Northcentral  Colorado.  Journal  of  Wildlife  Diseases  33:1–6.  Williams,  E.  S.,  and  S.  Young.  1980.  Chronic  wasting  disease  of  captive  mule  deer:  a  spongiform  encephalopathy.  Journal  of  Wildlife  Diseases  16:89–98.  Williams,  E.  S.,  and  S.  Young.  1992.  Spongiform  encephalopathies  of  Cervidae.  Revue  Scientifique  et  Technique  11:551–567.  Williams,  E.  S.,  M.  W.  Miller,  T.  J.  Kreeger,  K.  R.  H.  Kahn,  and  E.  T.  Thorne.  

30

30

2002.  Chronic  wasting  disease  of  deer  and  elk:  a  review  with  recommendations  for  management.  Journal  of  Wildlife  Management  66:551–563.  Williams,  E.  S.,  and  M.  W.  Miller.  2003.  Transmissible  spongiform  encephalopathies  in  non-­‐domestic  animals:  origin,  transmission  and  risk  factors.  Vol.  22,  pp.  145–156  in  Bengis,  R.  G.,  editor.  Risk  analysis  of  prion  diseases  in  animals.  Revue  scientifique  et  technique.  International  Office  of  Epizootics,  Paris,  France.  Wozencraft,  W.  C.  1978.  Investigations  concerning  a  high-­‐density  white-­‐tailed  deer  population  in  south  central  Wisconsin.  Thesis,  University  of  Wisconsin,  Influence of genetic relatedness and spatial proximity on chronic wasting disease infection

among female white-tailed deer

DANIEL A GREAR1, MICHAEL D SAMUEL2, KIM T SCRIBNER3, BYRON V

WECKWORTH4, JULIE A LANGENBERG5

1Department of Wildlife Ecology, University of Wisconsin, 1630 Linden Dr, Madison, WI 53706

2USGS Wisconsin Cooperative Wildlife Research Unit, University of Wisconsin, 1630 Linden Dr,

Madison, WI 53706

3Department of Fisheries and Wildlife and Department of Zoology, Michigan State University,

East Lansing, MI 48824

4Department of Biology, University of California, Riverside, CA 92521

5Wisconsin Department of Natural Resources, P.O. Box 7921, Madison, WI 53707-7921

Correspondence: Daniel A Grear, Center for Infectious Disease Dynamics, 208 Mueller Labs,

Pennsylvania State University, University Park, PA 16801, USA. E-mail: dag268@psu.edu

Running headline: Social transmission of CWD

Summary

1. We evaluated the hypothesis of socially facilitated transmission of chronic wasting disease

(CWD) among adult female white-tailed deer using spatial location and genetic relatedness for

31

31

1,387 female deer, as well as the spatial location of 1,321 adult male deer harvested during 2002-

2004 CWD control efforts in Wisconsin, USA.

2. There was little correlation between genetic relatedness and geographic distance among all

pairs of adult females separated by up to 10 km. At small distances (<0.8 km), CWD positive

deer were more related than random individuals in the population, indicating a weak association

between relatedness and CWD infection. This relationship was confounded by a low degree of

spatial aggregation of related females compared to previous theories.

3. Infection in adult females was strongly influenced by closely related females (full-sibling,

mother-offspring) that were spatially proximate (<3.2 km). To a lesser extent, infection was also

influenced by the number of infected nearby females (<3.2 km). In contrast, infection was not

influenced by less genetically related females (i.e. half-sibling, grandmother-granddaughter,

cousin) that were also expected to be in the same social group. Our results suggest a hierarchy of

CWD transmission within social groups based on familial relationships.

4. Our results indicate that direct (deer-to-deer) transmission of CWD likely occurs between

closely related female deer. CWD transmission also appears to occur among spatially proximate

females. However, we cannot determine whether occasional direct contact or contact with a

contaminated environment is responsible for increased infection among proximate females. It is

likely that direct and environmental (deer-environment-deer) transmission occur in this

epidemic.

5. The influence of spatially proximate females and close female kin on CWD infection does not

extend beyond a 2-3 km radius. In addition, the spatial relationship between infected females

and among infected males and females suggests that CWD transmission operates on a local scale

of 20-30 km2.

32

32

Key words: disease ecology, epidemiology, microsatellite genetic markers, social

structure, transmissible spongiform encephalopathy (TSE).

Introduction

Diseases are increasingly recognized as important processes in the ecology, population

dynamics, life history, and conservation of many vertebrate species (Hudson et al. 2002). As a

result, more attention is being given to the characteristics of host-parasite interactions that

influence transmission, including host social structure. Characteristics such as size and

demographic composition of social groups, inter-group movement, and inter-group isolation in

space and time influence the likelihood and duration of infectious contacts between individuals

as well as contact with contaminated environments (Freeland 1979; Altizer et al. 2003; Loehle

1994). In particular, social grouping plays a large role in disease transmission in primates

(Freeland 1979), ungulates (Cross et al. 2004), badgers (Delahay et al. 2000), and humans

(Anderson & May 1992). A key question in predicting disease spread is how social grouping

influences infectious contacts and determines density and frequency-dependent transmission

(McCallum, Barlow, & Hone 2001; Begon et al. 2002; Altizer et al. 2003). Pathogens

transmitted during social interactions (i.e. grooming, agonistic behavior) should spread more

quickly at higher population density or as groups become larger because infectious contacts

increase (Freeland 1976; Loehle 1995; Altizer et al. 2003). Alternatively, if social structure

restricts mixing of animals within a population or local area, pathogen spread is determined by

how often contacts occur (frequency-dependent transmission) (Smith et al. 1995; Altizer et al.

2003; Cross et al. 2004). Greater understanding of how sociality shapes infectious contact is

important for predicting disease spread and formulating effective management actions.

33

33

However, relatively few empirical studies addressing the effects of sociality on disease

transmission exist for wild mammal populations (Altizer et al. 2003).

Chronic wasting disease (CWD) is an emerging neurological disease affecting North

American species of deer (Odocoileus spp.), rocky mountain elk (Cervus elaphus L.), and moose

(Alces alces L.) and CWD is the only known transmissible spongiform encephalopathy (TSE)

that acts as an infectious disease in wildlife populations (Williams 2005). The infectious agent is

likely a prion, an abnormal form of a protein that naturally occurs in nervous tissue. CWD has a

lengthy infection period preceding clinical signs leading to eventual mortality (Williams and

Young 1980; Williams 2005). Studies of captive cervids have demonstrated that the CWD agent

can be transmitted via direct animal contact or from contact with an environment previously

occupied by an infected animal (Sigurdson et al. 1999; Miller & Williams 2003; Miller et al.

2004). The results of a recent inoculation experiment have demonstrated that oral inoculation

with saliva of infected deer was infective, while oral inoculations with feces or urine of infected

deer could not produce infection (Mathiason et al. 2006). However, no studies have been

conducted in free-ranging animals to determine the relative importance of direct deer-to-deer

transmission (referred to as direct transmission hereafter) and transmission from an infectious

deer to the environment to a susceptible deer (referred to as environmental transmission

hereafter). Thus, much remains unclear regarding the routes and rates of CWD transmission in

wild cervid populations (Gross & Miller 2001; Williams et al. 2002). There is no treatment for

CWD, but disease susceptibility appears to be affected by prion protein (PrP) genotype

(O’Rourke et al. 1999; Johnson et al. 2003; Johnson et al. 2006). Johnson et al. (2006) found

that susceptible genotypes occur in 91 - 98% of white-tailed deer (O. virginianus Zimmerman) in

34

34

our study area and concluded that there is no genetic barrier to CWD transmission due to PrP

genotype.

Gross & Miller (2001) proposed a frequency-dependent transmission model for CWD in

mule deer (O. hemionus Rafinesque) based on the type of social interactions that occur among

animals and the contact rate between social groups. However, the role of social groups and the

scale over which CWD transmission occurs have received little critical evaluation, but are crucial

to understanding and predicting CWD dynamics (Schauber & Woolf 2003; Farnsworth et al.

2006). Studies in mule deer (Miller & Conner 2005) and white-tailed deer (Grear et al. 2006)

have reported higher prevalence of CWD among males than among females along with higher

prevalence in adults than in fawns or yearlings. These results suggest transmission dynamics

may differ between the sexes due to differences in social behavior, as the host deer in these

studies have very different densities, live in contrasting habitats, but have similar social structure.

Similar male-biased infection patterns for Mycobacterium bovis (bovine tuberculosis) in white-

tailed deer from Michigan, USA (O’Brien et al. 2002), along with a significant relationship

between social structure and M. bovis infection (Blanchong et al. 2006b), further imply that

social behavior may be key to understanding infectious disease transmission in deer populations.

White-tailed deer social structure is based on segregation of sexes outside of the breeding

season and formation of female social groups outside the fawning period. Males typically have

larger home ranges and more social interactions than females, increasing their chance of

infectious contact with infected deer or with a contaminated environment (Hawkins & Klimstra

1970; Hirth 1976; Miller & Conner 2005; Grear et al. 2006). Female social behavior is centered

on matrilineal social units comprised of an older female, her daughters, and several generations

of closely related female offspring (Hawkins & Klimstra 1970). Successive generations of

35

35

females establish home ranges that overlap their natal range and kin (e.g., siblings, mother-

offspring, grandparent-grandchild) have extensive range overlap and social interaction (Porter et

al. 1991; Nelson & Mech 1999). Due to the strong philopatry and low dispersal of female white-

tailed deer, matrilineal groups are expected to be genetically related and spatially clustered

(Mathews & Porter 1993; Aycrigg & Porter 1997; Nelson & Mech 1999). Porter et al. (1991)

proposed a general model to describe this social structure called the ‘rose-petal hypothesis,’ as it

predicts that female social groups expand in a series of overlapping home ranges that appear

similar to the petals of a rose. However, female social groups are not territorial and maintain

varying degree of spatial overlap (Scribner et al. 1985; Porter et al. 1991; Mathews & Porter

1993; Aycrigg & Porter 1997; Comer et al. 2005; Skuldt 2005).

The ideal way to evaluate how social structure influences transmission is to quantify

inter-individual contact rates. Directly measuring contact rates in an entire wild deer population

is unrealistic and traditional methods of measuring social structure (i.e. direct observation, radio

telemetry) limit the number of individuals and groups that can be studied. However, recent

advances in multi-locus microstellite genetic markers (Queller, Strassman & Hughes 1993)

allowed us to determine the genetic relatedness among female deer removed from the CWD

infected area in Wisconsin, USA and employ this unique approach to estimate contact through

genetic relatedness (Scribner et al. 2005) and study infection patterns. We combined genetic

relatedness, spatial proximity, and demographic information to evaluate the probability of CWD

infection in individual adult female white-tailed deer in relation to age, kinship between females,

and the frequency and proximity of infected males and females in space.

The overall goal of our research was to evaluate the importance of social structure in the

transmission of CWD among adult female white-tailed deer. We focused on female deer

36

36

because their social philopatry predicts a strong spatial genetic pattern while high male dispersal

rates predict little correlation between social interactions and genetic patterns among bucks. In

addition, sexual segregation and differences in CWD prevalence between male and female

white-tailed deer suggest that separate transmission processes may be occurring (Miller &

Conner 2005, Grear et al. 2006). Based on the likelihood of both direct and environmental

transmission routes, we hypothesized that female white-tailed deer have greater CWD

transmission via direct contact or a shared contaminated environment among females in their

social group. Conversely, social barriers to contact between groups results in lower transmission

between members of different groups. This hypothesis predicts specific associations between

CWD infection, spatial proximity, and genetic relatedness: 1) infected females in close spatial

proximity will be more genetically related than the population average over the same area 2)

closely related females in the same social group are more likely to be infected than unrelated

deer in the local area, and 3) the probability of female infection is correlated more strongly with

local female infection than local male infection.

Materials and methods

STUDY AREA AND DATA COLLECTION

Following the discovery of CWD from white-tailed deer harvested in south-central

Wisconsin, USA during 2001, Joly et al. (2003) identified a 310 km2 area where prevalence was

higher (6.5%) than surrounding CWD management areas (1.5%) (Fig. 1). All deer examined in

this study were removed from this core study area. Prior to CWD management (post-harvest

2001) deer density in the core study area was estimated at 13.5 – 15.5 deer/ km2 (Rolley 2002).

As a result of generally temperate climate and abundant resources, white-tailed deer in this area

37

37

have high fecundity, exhibit very little seasonal movement, and have no seasonal migrations

(Larson, Rongstad, & Teilcox 1978; Ishmael 1984).

Deer were removed by hunter harvest and limited culling by government sharpshooters

from the core study area during April 2002-April 2003 and July 2003-January 2004 (Bartelt,

Pardee & Thide 2003). Age, sex, and kill location to the quarter section (65 ha) were recorded

for all deer harvested. Deer were assigned geographic coordinates as the center of the quarter-

section where harvested and this location was used as an estimate of the center of each deer’s

home range. Age was determined by tooth replacement and wear (Severinghaus 1949). Heads

were removed from harvested deer and sent to a tissue extraction center where a portion of the

brainstem (obex), retro pharyngeal lymph nodes (RPLN), and muscle tissue were collected for

CWD diagnosis and genetic analysis. At dissection, a portion of the obex and one RPLN were

immediately fixed in 10% buffered formalin and the remaining tissues were frozen. RPLN and

obex tissues were sent to the Wisconsin Veterinary Diagnostic Lab (Madison, Wisconsin, USA)

for CWD testing by immunohistochemistry (IHC) (Miller & Williams 2002) or plate ELISA

(IDEXX Laboratories Inc, Westbrook, Maine, USA). Fixed RPLN tissues from all deer

harvested in 2002 were tested using IHC. Positive RPLN tests were confirmed with IHC of the

obex. In 2003, most deer were screened with ELISA tests of RPLN and suspect positives were

confirmed using IHC of fixed RPLN and obex. A small number of deer harvested in 2003 were

only tested by IHC (D. Keane, Wisconsin Veterinary Diagnostic Laboratory, unpublished data).

For disease reporting and analysis, tests showing positive IHC results in the RPLN or the obex

were classified as positive for CWD.

GENOTYPES, RELATEDNESS, AND PEDIGREE RELATIONSHIPS

38

38

A 5-10g sample of muscle tissue from all adult females harvested from the core study

area during 2002-2004 was collected from frozen tissues. Muscle samples were stored in 95%

ethanol at -20o C until DNA extraction. DNA was isolated using a QAIGEN DNeasy spin-

column procedure (Qiagen Genomics Inc, Bothell, Washington, USA). Twelve polymorphic

microsatellite loci (Table 1) were chosen from a suite of loci previously optimized for white-

tailed deer (Kirkpatrick 1992; Bishop et al. 1994; DeWoody, Honeycutt, & Skow 1995; Talbot,

Haigh, & Plante 1996; Wilson et al. 1997; Blanchong et al. 2002) based on allelic diversity,

confirmation of Hardy-Weinberg equilibrium, and independence from other loci. DNA was

amplified using the polymerase chain reaction procedure and electrophoresed on 6%

polyacrylamide gels. Genotypes were visualized on a Hitachi FMBIOII gel scanner (MiraiBio

Inc., Alemeda, California, USA) or LI-COR gel scanner (LI-COR Biosciences, Lincoln,

Nebraska, USA) and scored manually.

Pairwise linear distances (dij) among all female pairs and all female-male pairs were

calculated using the program Passage v1.1 (Rosenberg 2001). Pairwise genetic distances (rxy)

among pairs of adult female deer were calculated using program Kinship v1.3.1 (Queller &

Goodnight 1989). Queller and Goodnight’s rxy is an unbiased estimate of relatedness based on

the population allele frequencies and ranges from -1 to 1. A value of 0 indicates that two

individuals are not more or less related than a pair drawn randomly from the population, a

positive value indicates a pair is more related, and a negative value indicates a pair is less related

than average.

Spatial genetic structure was examined by correlating pairwise genetic distance and

pairwise linear distance (r, Peakell & Smouse 2005) among all female deer and separately for

CWD infected females using a Mantel test for matrix correlation (Smouse, Long, & Sokal 1986).

39

39

The significance of the Mantel test statistic was determined based on 999 permutations using the

program GenALEx v 6 (Peakall & Smouse 2005). Geographic heterogeneity in genetic

relatedness was examined by measuring the mean correlation of genetic distance (r), which is

analogous to rxy, among female deer within 0.8 km distance classes using the program GenALEx

v 6 (Peakall & Smouse 2005). The program GenALEx estimated 95% confidence intervals (CI)

around r for each 0.8 km distance class by bootstrap resampling (n = 1,000) and created a 95%

confidence envelope from the hypothesis of no genetic structure in space by simulating a

population of random pairwise linear distances based on observed genotypes (n = 999).

The potential importance of social contact in CWD transmission was evaluated by

estimating specific pedigree relationships between female deer using likelihood ratio tests in the

program Kinship v 1.3.1 (Queller & Goodnight 1989). This method estimated type 1 error rates

(false positive) and type 2 error rates (false negative) for user defined primary and null pedigree

relationships by comparing observed genetic data to simulated pairs generated from observed

allele frequencies. Kinship v 1.3.1 was set to simulate 350,000 pairs to achieve a balance

between computing resources and estimating precise likelihood values for our primary and null

hypotheses. Pairs of female deer were identified that were close kin (i.e. full-sibling, mother-

offspring), called kin-class 1, based on a primary and null hypothesis of full-sibling (rxy = 0.5)

and half-sibling (rxy = 0.25), respectively. A second kinship class (kin-class 2) was identified

comprised of pairs of females that were also closely related, but with a lower degree of genetic

similarity based on primary and null hypotheses of full-sibling (rxy = 0.5) and not related (rxy =

0), respectively. All other pairs were considered unrelated. Females related in either kinship

class are expected to share social group membership based on the rose-petal hypothesis of female

social structure (Porter et al. 1991; Mathews & Porter 1993).

40

40

The proportion of co-infected pairs was estimated within each kinship class (kin-class 1,

kin-class 2, and unrelated) and a 95% CI was created using jackknife procedures. Jackknifed

estimates of variance were used because the sample size of unique pairs was artificially large

(961,191 pairs arose from 1,387 individuals) and was inappropriate for use in standard statistical

tests. If the 95% CI of co-infection probability of a kin-class did not overlap the mean

probability of co-infection, pairs in that kin-class were considered to have different risk of being

co-infected than average.

PAIRWISE CO-INFECTION AND SPATIAL DISTANCE

In order to assess the influence of spatial separation on transmission, a general non-linear

model was used to estimate the number of co-infected pairs as a function of linear distance

between individuals in each pair (dij):

co-infection = β1/(1+dij)β2 + β0 eqn 1

The total number of pairs at each distance was used as an offset to determine the prevalence of

co-infected pairs as a function of distance. This relationship was calculated for all female-female

pairs and separately for all female-male pairs. In equation 1, β0 represents the average

probability of pair co-infection, β1 describes the strength of the relationship between pairwise

spatial distance and pairwise infection, and β2 describes the shape of that relationship (i.e. if β2 =

-1 the relationship is linear, if β2 ≠ 1 the relationship is non-linear).

INDIVIDUAL INFECTION PROBABILITY

Multivariate models were used to evaluate whether factors related to potential exposure

time (age), the number and proximity of infected males and females, and potential social contact

(kinship) of infected females were associated with risk of CWD infection in adult female deer  

41

41

Potential disease risk associated with CWD positive deer in geographic space was

estimated by calculating variables based on the number of nearby infected females or males for

each female, i:

where pj is one if individual j is CWD positive and 0 otherwise, fj is one if individual j is female

and 0 otherwise, mj is one if individual j is male and 0 otherwise, and dij is the spatial distance

between individual i and j. No risk was assigned if a CWD positive deer was harvested beyond

3.2 km (2 mi); the distance where the probability of two females or males being co-infected

approaches the population average (Fig. 3) and regular contact or range overlap are unlikely

(Skuldt 2005).

Parameters estimating risk based on proximity to infected individuals were calculated as

the sum of the inverse distances to infected females or males:

n Σ j=1,

j≠i

(pj)(fj)(1+dij)-1 if dij < 3.2 km 0 if dij > 3.2 km

dposFemalei = eqn 4

n Σ j=1,

j≠i

(pj)(mj)(1+dij)-1 if dij < 3.2 km 0 if dij > 3.2 km

dposMalei = eqn 5

(pj)(fj) if dij < 3.2 km 0 if dij > 3.2 km

n Σ j=1,

j≠i

nposFemalei = eqn 2

n Σ j=1,

j≠i

(pj)(mj) if dij < 3.2 km 0 if dij > 3.2 km

nposMalei = eqn 3

42

42

Risk parameters were also created for each adult female based on her kinship with CWD

positive females. One set of covariates represented potential risk of CWD infection for each

adult female based on her kinship to other CWD positive adult females that were separated by

3.2 km or less:

A second set of covariates based on CWD positive kin that were harvested at distances

greater than 3.2 km was calculated to evaluate whether it was necessary for infected kin to be

close enough to have a reasonable chance for regular contact to influence transmission:

where, pj is one if individual j is CWD positive and 0 otherwise, k1j is one if deer i and deer j are

related in kin-class 1, k2 j is one if deer i and deer j are related in kin-class 2, and dij is the linear

distance between i and j.

n Σ j=1,

j≠i

(pj)(k1j) if dij > 3.2 km 0 if dij < 3.2 km

kin1fari = eqn 8

n Σ j=1,

j≠i

(pj)(k2j) if dij > 3.2 km 0 if dij < 3.2 km

kin2fari = eqn 9

n Σ j=1,

j≠i

(pj)(k2j) if dij < 3.2 km 0 if dij > 3.2 km

kin2neari = eqn 7

n Σ j=1,

j≠i

(pj)(k1j) if dij < 3.2 km 0 if dij > 3.2 km

kin1neari = eqn 6

43

43

A series of logistic regression models was fit predicting the infection status (positive or

negative) of each adult female deer using the covariates representing spatial and genetic risk of

infection, and deer age. A variance inflation factor (ĉ) for our saturated model was fit to

determine whether overdispersion was present (Cox & Snell 1989). Information theory was used

to determine the relative weight of the competing models, to calculate model averaged

coefficients, and to make model averaged predictions based on the top models (Burnham &

Anderson 2002).

Results

The analysis included 1,387 adult females with disease status, location, and genotype

results. Of these, 77 (5.5%) were CWD positive. There were 1,321 adult males with location

and disease status, with 99 (7.5%) testing CWD positive. Total observed female and male

prevalence was 5.4% and 7.4%, respectively, over the same sampling period with prevalence

significantly higher in males than in females (Grear et al. 2006).

CORRELATION BETWEEN GENETIC RELATEDNESS AND SPATIAL DISTANCE

Pairwise linear distances were not correlated with pairwise genetic distances (r) for adult

females (Mantel test for matrix correlation = 0.109, randomization p = 0.22, n = 999), nor was

there a correlation between genetic distance and linear distance among the CWD positive adult

females (correlation = 0.032, randomization p = 0.143, n = 999).

There was higher genetic relatedness than random among deer harvested at 3.2 km (2 mi)

or less with the level of relatedness declining from the smallest distance interval (Fig. 2a). A

similar analysis of CWD positive females showed that genetic relatedness was highest at the

smallest distance class, but not statistically different from random (Fig. 2b). Large variances in

the relatedness of CWD positive females along with a small number of CWD infected females

44

44

resulted in genetic relatedness that was not significantly different from the population average at

any distance (Fig. 2a-b).

KINSHIP INFECTION

Four of 182 (0.022) pairs were co-infected in kin-class 1, 6 of 1,105 (0.005) pairs were

co-infected in kin-class 2, and 2,916 of 959,904 (0.003) pairs were co-infected in the unrelated

class (Table 1). Deer in kin-class 1 were co-infected at a significantly (95% jackknife CI =

[0.014 , 0.030]) higher rate than random (0.0031), but deer in kin-class 2 (95% jackknife CI =

[0.0016 , 0.009]) or unrelated deer (95% jackknife CI = [0.0001 , 0.006]) were not.

CO-INFECTION PROBABILITY AND PAIRWISE DISTANCE

There was a decline in the probability of co-infection with distance in both female-female

and female-male pairs (Fig. 3 a-b). Binomial confidence intervals for the proportion of co-

infected pairs approached the population mean as distances between individuals reached 2-4 km

for both female-female and female-male pairs (Fig. 3 a-b). Coefficients from equation 1 for

female-female pairs were: β0 = 0.0013 (95% CI [-0.00003, 0.0027]), β1 = 0.0053 (95% CI

[0.0043, 0.0064]), and β2 = 0.6991 (95% CI [0.2128, 1.185]). Coefficients for female-male pairs

were: β0 = 0.0022 (95% CI [0.0012, 0.0033]), β1 = 0.0063 (95% CI [0.0055, 0.0072]), and β2 =

0.7478 (95% CI [0.3800, 1.1156]). The decline in prevalence with distance appears to be non-

linear, but the 95% CI of β2 for female-female pairs and female-male pairs overlaps 1 indicating

the inverse of distance is an adequate measure of the decline in co-infection given the resolution

of our data. These results support our use of an inverse linear relationship between distance and

infection risk in the logistic regression models predicting individual female infection (equations

4-5).

INDIVIDUAL FEMALE INFECTION PROBABILITY

45

45

Multivariate logistic regression models were evaluated to determine the importance of

age, infected female kin, and the number and proximity of CWD infected deer (males and

females) on the probability of infection for an adult female. There was no evidence that our data

was overdispersed based on the saturated model (ĉ = 0.40, parameters = 10). Therefore no

corrections for overdispersion were used for model selection or parameter variances. There was

no single best model that predicted individual infection and five models had ΔAIC < 2, while no

other model had ΔAIC < 2.6 (Table 3). These five models accounted for a total of 65% of the

Aikaike weights and were used to obtain model averaged covariate estimates and model

averaged predictions. After model averaging, a positive significant relationship was found

between the probability of infection and the number of nearby infected females (nposFemale =

0.08, 95% model averaged CI [ 0.02 , 0.04 ]) and the number of nearby infected deer in kin-class

1 (kin1near = 3.21, 95% model averaged CI [ 0.70 , 1.84 ]). The age of a female (age = 0.02,

95% model averaged CI [ -0.11 , 0.16 ], the covariate for the proximity of nearby infected

females (dposFemale = -0.29, 95% model averaged CI [ -1.15 , 0.57 ], the number of nearby

infected males (nposMale = -0.0002, 95% model averaged CI [ -0.005 , 0.005 ], the number of

nearby infected kin related in kin-class 2 (kin2near = -0.08, 95% model averaged CI [ -0.31 ,

0.16 ], the number of distant infected kin related in kin-class 2 (kin2far = 0.08, [ -0.02 , 0.19 ],

and the number of distant infected kin related in kin-class 1 (kin1far = 1.31, 95% model averaged

CI [ -0.01 , 2.63 ] were present in the top models, but did not have parameter estimates

significantly different from zero.

To illustrate the relative effects of the significant model parameters, model averaged

predictions for the probability of infection in an adult female were compared for an increasing

number of infected females (nposFemale) with zero and one infected deer in kin-class 1 located

46

46

nearby (kin1near) (Fig. 4). Probability of infection for an average aged female deer (2.89 yrs)

with average values of the non-significant parameters (dposFemale, nposMale, kin2near,

kin2far, kin1far) increased slightly with more nearby infected females (nposFemale) (Fig. 4). A

CWD positive deer related in kin-class 1 was rare in our data-set and no female had more than

one CWD positive kin. However, the presence of a single CWD positive female in kin-class 1

dramatically increased the probability of infection (Fig. 4).

Discussion

SOCIAL INTERACTIONS INFLUENCE TRANSMISSION

Genetic relatedness and proximity among infected females was used to evaluate whether

social group affiliation increased the probability of CWD infection in adult female deer. A

stronger association with genetic relatedness than spatial distance was predicted because frequent

and intimate contacts occur more often among females in the same social group than among

females in different social groups (Hirth 1976; Schauber et al. 2006). There was higher, but not

statistically different relatedness (r) than expected among spatially close CWD positive female

deer (< 0.8 km, Fig. 2b), but not at distances greater than 0.8 km (Fig. 2b). This weak correlation

between genetic relatedness and linear distance among CWD positive females provides limited

support for the hypothesis of increased transmission within social groups. However, our ability

to detect relationships between relatedness (e.g. social affiliation) and disease status was

confounded by a low degree of genetic structure among female deer with close spatial

associations (Fig. 2a), which is contradictory to the prediction of strong female social structure

(Porter et al. 1991). In contrast to our results, studies of white-tailed deer genetic structure on

similar geographic scales reported much higher levels of relatedness, and were able to

47

47

demonstrate a clear relationship between greater infection with M. bovis and higher genetic

relatedness (Blanchong et al 2006a; Blanchong et al. 2006b).

When specific kin relationships were considered, kin-class 1 pairs were more likely to be

co-infected than either kin-class 2 relatives or unrelated females (Table 2). Further, our

multivariate logistic regression analysis indicated the number of infected females within 3.2 km

and infection in close female kin were significantly associated with CWD infection in adult

female deer, with a greater than 10-fold increase (from 4% to 52%) with one infected female kin

nearby (Fig. 4). This level of genetic relatedness corresponded to full-sibling or mother-daughter

pairs (Queller & Goodnight 1989), which likely have intense social contacts and high spatial

overlap (Hirth 1976; Nelson & Mech 1999; Schauber et al. 2006). These results suggest that

social behavior among close relatives strongly increases CWD transmission compared to

transmission among all females in a local area.

In contrast, infected kin-class 2 relatives had no influence on infection. Deer related at

kin-class 2 are likely to share social group membership, experience social interactions, and have

overlapping space use (Hirth 1976; Nixon 1991). Schauber et al. (2006) reported that direct

contact or extreme overlap in space use increased nearly 20-fold among female white-tailed deer

that were members of the same social group compared to other females that occurred in the same

area. While social interaction, as measured by genetic relatedness, appears to increase CWD

infection, transmission does not appear to be uniform among all members of a social group.

Alternative mechanisms may also be responsible for higher infection among the most

closely related females because closely related females experience similar contacts with other

deer (males, unrelated females) and have similar space use. Similar PrP genetic sequence has the

48

48

potential to influence infection among kin, however, it is unlikely to affect our analyses because

nearly all deer in our study area have fully susceptible PrP genotypes (Johnson et al. 2006).

INFECTION PATTERNS SUGGEST THAT DIRECT AND ENVIRONMENTAL

TRANSMISSION MODES OCCUR

The evidence for a social influence on CWD transmission can also provide insights into

the route of infection in natural deer populations. The strong influence of infected kin-class 1

relatives along with the lack of influence of infected kin-class 2 relatives on both pairwise

infection (Table 2) and individual infection (Table 3) indicates that direct transmission between

kin-class 1 females is responsible for increased infection, as less related females are predicted to

be in the same social group and have similar range use, but did not influence infection. The

mechanism for transmission between the most closely related females is likely contact during

social behaviors. Recent inoculation trials that demonstrated saliva was infective, while feces

and urine were not, highlight the potential of direct transmission during social contact

(Mathiason et al. 2006). Although male infection did not influence the probability of infection in

spatially proximate females (Table 3), we can not rule out the possibility that closely related

pairs became infected from a common source, because they experience similar contact with other

deer (males, unrelated females) and have similar space use.

Our results also demonstrated a positive correlation between probability of infection and

the number of infected nearby females and this relationship was much weaker than the

relationship we found between infected kin-class 1 females and probability of infection (Fig. 4).

However, we are not able to determine the mechanism responsible for this association. It seems

likely that environmental transmission may be responsible for CWD spread among these

49

49

spatially proximate female deer because they share varying degrees of social interactions, which

suggests heterogeneous and low probabilities of direct contact.

If CWD transmission is primarily driven by direct contact, then contact among members

of the same social group is the most likely route for transmission between females in the same

area. A consequence of this scenario is that CWD transmission rates among females may be

frequency-dependent, because direct contact between unrelated females is limited, even if social

groups have extensive range overlap. Transmission may increase within a social group as group

size increased, but spread beyond social groups would not, as long as groups have segregated

contact.

Alternatively, if environmental transmission were primarily responsible for CWD spread,

we believe social structure would have a small influence on transmission because susceptible

deer that overlap spatially could acquire infection regardless of social affiliation or sex. A

consequence of this scenario is that CWD transmission rates among females would be density-

dependent because environmental sources of infection would be available to any female

regardless of social affiliation. Transmission would increase as population density increased if

social group size increased, the number of overlapping social groups increased, or the extent of

range overlap among social groups increased because more individuals would have the potential

to contact a common infectious source.

Given the increased infection among kin-class 1 relatives, the smaller positive influence

on infection due to nearby females, and the apparent transmission potential of saliva (Mathiason

et al. 2006), we believe that direct transmission is a key process involved in CWD transmission

among females at the scale of our study, with a lower level of environmental transmission

occurring concomitantly. However, we also note that the number of spatially proximate males

50

50

(nposMale) and females (nposFemale) were correlated (correlation coefficient, r = 0.62),

emphasizing the need for continued research into the source of infectious prion material in

natural settings, the mechanisms for environmental transmission, and the potential transmission

between male and female deer. In a broader sense, given the overall weak genetic structure in

our white-tailed deer, the fact that significant correlations between related deer and CWD

infection were detected, strongly suggests that transmission models should consider social

structure.

We recognize potential challenges in detecting fine-scale genetic structure in natural

populations with the available genetic methods. There are no clear guidelines to determine the

optimal number of loci for natural populations and it is possible that more than 12 loci could

have provided greater accuracy for genetic relatedness and kinship estimation in our study. For

example, the estimation of full-sibling pedigree relationships can usually be improved with more

genetic information. However, Butler et al. (2004) found that using 8 loci with 8 alleles each

was sufficient for accurate pedigree and full-sibling classification using several algorithms,

including maximum likelihood kinship methods similar to those we employed. In addition, we

used stringent criteria to ensure that pairs related in kin-class 1 share enough genetic information

to be considered full-sibling, but also share more genetic information than expected for lower

pedigree relationships (kin hypotheses, Table 2, see Queller & Goodnight 1989 for details on

kinship hypothesis testing).

SPATIAL RANGE OF CWD TRANSMISSION

Our results indicate that CWD is likely transmitted among adult female deer within a

local area of approximately 20-30 km2 (2.5 – 3 km radius) or less. There was no correlation

between the probability of infection and number of infected deer farther than 3.2 km (Fig. 3). In

51

51

addition, there was no influence of infected kin if they were located farther than 3.2 km,

indicating that distance between females serves as a barrier to transmission.

Previous spatial analyses of CWD prevalence in white-tailed deer also identified spatial

autocorrelation patterns of infection at a scale (<3.6 km) similar to our results (Joly et al. 2006).

Proximity to other infected female deer had a positive influence on CWD transmission, but

determining the underlying mechanism is difficult due to the uncertainties regarding the relative

importance of direct and environmental transmission. Both routes can be related to spatial

proximity and, due to the high deer density and harvest pressure in southern Wisconsin, there is

likely considerable overlap of female social groups. High spatial overlap among social groups

may be one of the reasons we do not find a stronger overall association between genetic

relatedness (r) and CWD infection or observe spatial infection patterns consistent with highly

structured social groups, like those observed for M. bovis infection in Eurasian badgers (Meles

meles) (Delehay et al. 2000) or for M. bovis infection in white-tailed deer (Blanchong et al.

2006b).

RESEARCH NEEDS

CWD is a newly emerging disease in North American cervid populations and much of the

basic science concerning disease transmission and spread is not well understood. Two important

areas for future research are the relative importance of direct and environmental transmission and

the routes of transmission among wild cervids. Research that illuminates the relative importance

of these two transmission modes and what excreta, discharges, or tissues contains infectious

materials under natural conditions will greatly improve our interpretation of epidemiology

studies, our efforts to model CWD spread, and our ability to formulate management actions.

Research that demonstrates the mechanisms driving higher prevalence in males and the

52

52

importance of transmission among males and females will also be important to understanding

and managing CWD outbreaks.

Acknowledgements

This study could not have been completed without the time and hard work of the

Wisconsin Department of Natural Resources staff and volunteers as well as the Wisconsin

Veterinary Diagnostic Lab TSE testing staff. We thank Carl Batha, Tom Howard, Matt Watrud,

Kerry Beheler, and Janet Sausen of the WDNR and Phillip Boschler and Delwyn Keane of the

Wisconsin Veterinary Diagnostic Laboratory for the time and effort they put into sample

collection, data management, and disease testing. We thank Kristi Filcek, Laura Main, and Anna

Simon for their expertise and hard work in the Molecular Ecology Laboratory at Michigan State

University. Funding was provided through the United States Geological Survey National

Wildlife Health Center and the Wisconsin Department of Natural Resources. We also

acknowledge David Coltman, Mary Conner, and Eric Schauber for valuable comments on early

drafts of this manuscript.

References

Altizer, S., Nunn, C.L., Thrall, P.H., Gittleman, J.L., Antonovics, J., Cunningham, A.A.,

Dobson, A.P., Ezenwa, V., Jones, K.E., Pederson, A.B., Poss, M. & Pulliam, J.R.C. (2003)

Social organization and parasite risk in mammals: integrating theory and empirical studies.

Annual Review of Ecology, Evolution, and Systematics, 34, 517-547.

Anderson, R.M. & May, R.M. (1992) Infectious diseases of humans: dynamics and control.

Oxford University Press, New York.

Aycrigg, J.L. & Porter, W.F. (1997) Sociospatial dynamics of white-tailed deer in the central

Adirondack Mountains, New York. Journal of Mammalogy, 78, 468-482.

53

53

Bartelt, G., Pardee, J, & Thiede, K. (2003) Environmental impact statement on rules to eradicate

chronic wasting disease in Wisconsin’s free-ranging white-tailed deer herd. Wisconsin

Department of Natural Resources, Madison, Wisconsin, USA.

Begon, M., Bennett, M., Bowers, R.G., French, N.P., Hazel, S.M. & Turner, J. (2002) A

clarification of transmission terms in host-microparasite models: number, densities and

areas. Epidemiology and Infection, 129, 147-153.

Bishop, M.P., Kappes, S.M., Keele, J.W., Stone, R.T., Sunden, S.F., Hawkins, G.A., Toldo, S.S.,

Fries, R., Grosz, M.D., Yoo, J. & Beattie, C.W. (1994) A genetic linkage map for cattle.

Genetics, 136, 619-639.

Blanchong, J.A., Scribner, K.T. & Winterstein, S.R. (2002) Assignment of individuals to

populations: Bayesian methods and multi-locus genotypes. Journal of Wildlife

Management, 66, 321-329.

Blanchong, J.A., Scribner, K.T., Epperson, B.K. & Winterstein, S.R. (2006a) Changes in

artificial feeding regulations impact white-tailed deer fine-scale spatial genetic structure.

Journal of Wildlife Management, 70, 1037-1043.

Blanchong, J.A., Scribner, K.T., Kravchenko, A.N. & Winterstein, S.R. (2006b) TB-infected

deer are more closely related than non-infected deer. Biology Letters, in press.

Burnham, K.P. & Anderson, D.R. (2002) Model selection and inference: a practical

information-theoretic approach. 2nd edition, Spring-Verlag, New York.

Butler, K., Field, C., Herbinger, C.M. & Smith, B.R. (2004) Accuracy, efficiency and robustness

of four algorithms allowing full sibship reconstruction from DNA marker data. Molecular

Ecology, 13, 1589-1600.

54

54

Comer, C.E., Kilgo, J.C., D’Angelo, G.J., Glenn, T.C. & Miller, K.V. (2005) Fine-scale genetic

structure and social organization in female white-tailed deer. Journal of Wildlife

Management, 69, 332-344.

Cox, D.R. & Snell, E.J. (1989) Analysis of binary data. 2nd edition, Chapman and Hall, New

York.

Cross, P.C., Loyd-Smith, J.O., Bowers, J.A., Hay, C.T., Hofmeyr, M. & Getz, W.M. (2004)

Integrating association data and disease dynamics in a social ungulate: bovine tuberculosis

in African buffalo in the Kruger National Park. Annales Zoologici Fennici, 41, 879-892.

Delahay, R.J., Langton, S., Smith, G.C., Clifton-Hadley, R.S. & Cheeseman, C.L. (2000) The

spatio-temporal distribution of Mycobacterium bovis (bovine tuberculosis) infection in a

high-density badger population. Journal of Animal Ecology, 69, 428-441.

DeWoody, J.A., Honeycutt, R.L. & Skow, L.C. (1995) Microsatellite markers in white-tailed

deer. Journal of Heredity, 86, 317-319.

Farnsworth, M.L., Hoeting, J.A., Hobbs, N.T. & Miller, M.W. (2006) Linking chronic wasting

disease to mule deer movement scales: a hierarchical Bayesian approach. Ecological

Applications, 16, 1026-1036.

Freeland, W.J. (1979) Primate social groups as biological islands. Ecology, 60, 719-728.

Grear, D.A., Samuel, M.D., Langenberg, J.A. & Keane, D. (2006) Demographic patterns and

harvest vulnerability of CWD infected white-tailed deer in Wisconsin. Journal of Wildlife

Management, 70, 546-553.

Gross, J.E. & Miller, M.W. (2001) Chronic wasting disease in mule deer: disease dynamics and

control. Journal of Wildlife Management, 65, 205-215.

55

55

Hawkins, R.E. & Klimstra, W.D. (1970) A preliminary study of the social organization of white-

tailed deer. Journal of Wildlife Management, 34, 407-419.

Hirth, D.H. (1976) Social behavior of white-tailed deer in relation to habitat. Wildlife

Monographs, 53, 1-55.

Hudson, P.J., Rizzoli, A., Grenfell, B.T., Heesterbeek, H. & Dobson, A.P. (2002) The ecology of

wildlife diseases. Oxford University Press, Oxford.

Ishmael, W.E. (1984) White-tailed deer ecology and management in Southern Wisconsin. M.S.

Thesis, University of Wisconsin-Madison, Madison, Wisconsin, USA.

Johnson, C., Johnson, J., Clayton, M., McKenzie, D. & Aiken, J. (2003) Prion protein gene

heterogeneity in free-ranging white-tailed deer within the chronic wasting disease affected

region of Wisconsin. Journal of Wildlife Diseases, 39, 576-581.

Johnson, C., Johnson, J., Vanderloo, J.P, Keane, D., Aiken, J.M. & McKenzie, D. (2006) Prion

protein polymorphisms in white-tailed deer influence susceptibility to chronic wasting

disease. Journal of General Virology, 87, 2109-2114.

Joly, D.O., Samuel, M.D., Langenberg, J.A., Blanchong, J.A., Batha, C.A., Rolley, R.E., Keane,

D.P. & Ribic, C.A. (2006) Spatial epidemiology of chronic wasting disease in Wisconsin

white-tailed deer. Journal of Wildlife Diseases, 42, 578-588.

Joly, D.O., Ribic, C.A., Langenberg, J.A., Beheler, K., Batha, C.A., Dhuey, B.J., Rolley, R.E.,

Bartelt, G., Van Deelen, T.R. & Samuel, M.D. (2003) Chronic wasting disease in free-

ranging Wisconsin white-tailed deer. Emerging Infectious Diseases, 9, 599-601.

Kirkpatrick, B.W. (1992) Identification of a conserved microsatellite site in the porcine and

bovine insulin-like growth factor-I gene 5’ flank. Animal Genetics, 23, 543-548.

56

56

Larson, T.J, Rongstad, O.J. & Tebilcox, F.W. (1978) Movement and habitat use of white-tailed

deer in south central Wisconsin. Journal of Wildlife Management, 42, 113-117.

Loehle, C. (1995) Social barriers to pathogen transmission in wild animal populations. Ecology,

76, 326-335.

Mathews, N.E. & Porter, W.F. (1993) Effect of social structure on genetic structure of free-

ranging white-tailed deer in the Adirondack Mountains. Journal of Mammalogy, 74, 33-

43.

Mathiason, C.K., Powers, J.G., Dahmes, S.J., Osborn, D.A., Miller, K.V., Warren, R.J., Mason,

G.L., Hays, S.A., Hayes-Klug, J., Seelig, D.M., Wild, M.A., Wolfe, L.L., Spraker, T.R.,

Miller, M.W., Sigurdson, C.J., Telling, G.C. & Hoover, E.A. (2006) Infectious prions in the

saliva and blood of deer with chronic wasting disease. Science, 314, 133-136.

McCallum, H., Barlow, N. & Hone, J. (2001) How should pathogen transmission be modeled?

Trends in Ecology and Evolution, 16, 295-300.

Miller, M.W. & Conner, M.M. (2005) Epidemiology of chronic wasting disease in free-ranging

mule deer: spatial, temporal, and demographic influences on observed prevalence. Journal

of Wildlife Diseases, 41, 275-290.

Miller, M.W. & Williams, E.S. (2003) Horizontal prion transfer in mule deer. Nature, 425, 35-

36.

Miller, M.W. & Williams, E.S. (2002) Detection of PrPcwd in mule deer by

immunohistochemistry of lymphoid tissues. Veterinary Record, 151, 610-612.

Miller, M.W., Williams, E.S., Hobbs, N.T. & Wolfe, L.L. (2004) Environmental sources of prion

transmission in mule deer. Emerging Infectious Diseases, 10, 1003-1006.

57

57

Nelson, M.E. & Mech, L.D. (1999) Twenty-year home range dynamics of a white-tailed deer

matriline. Canadian Journal of Zoology, 77, 1128-1135.

Nixon, C.M., Hansen, L.P., Brewer, P.A. & Chelsvig, J.E. (1991) Ecology of white-tailed deer in

an intensively farmed region of Illinois. Wildlife Monographs, 118, 1-77.

O’Brien, D.J., Schmitt, S.M., Fierke, J.S., Hogle, S.A., Winterstein, S.R., Cooley, T.M., Moritz,

W.E., Diegel, K.L., Fitzgerald, S.D., Berry, D.E. & Kaneene, J.B. (2002) Epidemiology of

Mycobacterium bovis in free ranging white-tailed deer, Michigan, USA, 1995-2000.

Preventive Veterinary Medicine, 54, 47-63.

O’Rourke, K.I., Besser, T.E., Miller, M.W., Cline, T.F., Spraker, T.R., Jenny, A.L., Wild, M.A.,

Zebarth, G.L., Williams, E.S. & Zebarth, L. (1999) PrP genotypes of captive and free-

ranging Rocky Mountain elk (Cervus elaphus nelsoni) with chronic wasting disease.

Journal of General Virology, 80, 2765-2769.

Peakall, R. & Smouse, P.E. (2005) GenALEx 6: Genetic Analysis in Excel. Population genetic

software for teaching and research. Australian National University, Canberra, Australia.

http://www.anu.edu.au/BoZo/GenAlEx/

Porter, W.F., Mathews, N.E., Underwood, H.B., Sage, R.W. & Behrend, D.F. (1991) Social

organization in deer: implications for localized management. Environmental Management,

15, 809-814.

Queller, D.C. & Goodnight, K.F. (1989) Estimating relatedness using genetic markers.

Evolution, 43, 258-275.

Queller, D.C., Strassmann, J.E & Hughes, C.R. (1993) Microsatellites and kinship. Trends in

Ecology and Evolution, 8, 285-88.

Rice, W.R. (1989) Analyzing tables of statistical tests. Evolution, 43, 223-225.

58

58

Rolley, R. (2002) White-tailed deer population status 2002. Wisconsin Wildlife Surveys

(Compiled by B. Dhuey). Wisconsin Department of Natural Resources. Monona,

Wisconsin, USA.

Rosenberg, M.S. (2001) PASSAGE. Pattern analysis, spatial statistics and geographic exegesis.

Version 1.1. Department of Biology, Arizona State University, Tempe, Arizona, USA.

http://www.passagesoftware.net/

Schauber, E.M., Storm, D.J. & Nielson, C.K. (2006) Effects of joint space use and group

membership on contact rates among white-tailed deer. Journal of Wildlife Management, in

press.

Schauber, E.M. & Woolf, A. (2003) Chronic wasting disease in deer and elk: A critique of

current models and their application. Wildlife Society Bulletin, 31, 610-616.

Scribner, K.T., Wooten, M.C., Smith, M.H. & Jones, P.E. (1985) Demographic and genetic

characteristics of white-tailed deer populations subjected to still or dog hunting. Game

Harvest Management (eds. S.L. Beason & S.F. Robinsin), pp. 197-212. Ceasar Kleberg

Wildlife Research Institute, Kingsville, Texas, USA.

Scribner, K.T., Blanchong, J.A., Bruggeman, D.J., Epperson, B.K., Lee, C.Y., Pan, Y.W.,

Shorey, R.I., Prince, H.H., Winterstein, S.R. & Luukkonen, D.R. (2005) Geographical

genetics: conceptual foundations and empirical applications of spatial genetic data in

wildlife management. Journal of Wildlife Management, 69, 1434-1453.

Severinghaus, C.W. (1949) Tooth development and wear as criteria of age in white-tailed deer.

Journal of Wildlife Management, 13, 195-216.

59

59

Sigurdson, C.J., Williams, E.S., Miller, M.W., Spraker, T.R., O’Rourke, K.L. & Hoover, E.A.

(1999) Oral transmission and early lymphoid tropism of chronic wasting disease PrPres in

mule deer fawns (Odocoileus hemionus). Journal of General Virology, 80, 2757-2764.

Skuldt, L.H. (2005) White-tailed deer habitat use and behavior in the chronic wasting disease

eradication zone in Wisconsin. M.S. Thesis, University of Wisconsin, Madison,

Wisconsin, USA.

Smith, G.C., Richards, M.S., Clifton-Hadley, R.S. & Cheeseman, C.L. (1995) Modeling bovine

tuberculosis in badgers in England: preliminary results. Mammalia, 59, 619-631.

Smouse, P.E., Long, J.C. & Sokal, R.R. (1986) Multiple regression and correlation extensions of

the Mantel test of matrix correspondence. Systematic Zoology, 35, 627-769.

Talbot, J., Haigh, J. & Plante, Y. (1996) A parentage evaluation test in North American elk

(wapati) using microsatellites of ovine and bovine origin. Animal Genetics, 27, 117-119.

Williams, E.S. & Young, S. (1980) Chronic wasting disease of captive mule deer: a spongiform

encephalopathy. Journal of Wildlife Disease, 16, 465-471.

Williams, E.S., Miller, M.W., Spraker, T.R., O’Rourke, K.L. & Hoover, E.A. (2002) Chronic

wasting disease of deer and elk: a review with recommendations for management. Journal

of Wildlife Management, 66, 551-563.

Williams, E.S. (2005) Chronic wasting disease. Veterinary Pathology, 42, 530-549.

Wilson, G.A., Stroneck, C., Wu, L. & Coffin, J.W. (1997) Characterization of microsatellite loci

in caribou, Rangifer tarandus, and their use in other artiodactyls. Molecular Ecology, 6,

697-699.

60

60

Table 1. Locus name, number of alleles, observed heterozygosity, expected heterozygosity, and

P-value for Hardy-Weinberg equilibrium (HWE) for 12 microsatellite loci genotyped in 1,387

female white-tailed deer harvested in southern Wisconsin, April 2002–January 2004.

Heterozygosity Locus Alleles Observed Expected HWE p-value 1225 12 0.706 0.745 <0.001a 4107 16 0.810 0.810 0.005a 4208 19 0.838 0.901 0.002a 6506 13 0.703 0.871 <0.001a IGF1 13 0.654 0.665 0.528 RT27 19 0.826 0.832 0.708 RT7 18 0.874 0.876 0.245 C1 17 0.774 0.768 0.113 C2 11 0.645 0.829 <0.001a

CSN3 6 0.328 0.477 <0.001a RT23 19 0.913 0.912 0.694 RT9 10 0.816 0.800 0.009

a Indicates locus not in Hardy-Weinberg equilibrium after sequential Bonferroni correction (Rice 1989)