Abstract

Mexican wolves (Canis lupus baileyi), classified as probably extinct in the wild in Mexico and endangered in the US, were reintroduced into Arizona in 1998. We combined annual serologic testing results from samples collected between 2003 and 2016 from 108 wolves and known survival data from 118 wolves born in the recovery area from 2003 to 2014 to evaluate whether exposure to canine distemper virus (CDV) or canine parvovirus (CPV) was associated with a greater risk of mortality before 2 yr of age. We used mixed-effects logistic regression to estimate the effect of CDV and CPV on the probability of mortality. Annual seroprevalence rates for CDV and CPV ranged from 0% to 62% and from 33% to 100%, respectively (median, 14.2% and 90.3%, respectively). The covariate, age at testing, had a negative effect on mortality, indicating that younger animals had lower survival, whereas sex had little effect on mortality. The best-supported model excluded any effect of CPV or CDV on death before 2 yr old at both the pack and individual level. Although our analysis did not detect an effect of these viruses on mortality before 2 yr old, CDV was later identified as the cause of mortality in two individuals in 2017. Additional information is needed to assess the impact of these diseases on Mexican wolves.

The Mexican wolf (Canis lupus baileyi) has been classified as endangered by Mexico (probably extinct in the wild) and the US. In Arizona, Mexican wolves were extirpated by the 1970s, and the first recovery plan was completed in 1982. In 1998, 11 captive-born wolves were released, and the first pups were born to wild-born parents in 2002 (Arizona Game and Fish Department 2017). The population increased to a minimum of 113 wolves in Arizona and New Mexico by the end of 2016. Mexican wolves have an average pack size of 4.1 animals (range, 2–14) consisting of an alpha male and female, plus one or more pups and adults (US Fish and Wildlife Service 2017). Litters are born in April or May and disperse from their natal pack between 9 mo and 36 mo.

Canine distemper virus (CDV) is a morbillivirus (Paramyxoviridae) that affects all terrestrial carnivores and has been associated with local population declines in some species, including the African lion (Panthera leo; Deem et al. 2000), Channel Island kit fox (Urocyon littoralis catalinae; Deem et al. 2000), and black-footed ferret (Mustela nigripes; Williams et al. 1988). Transmission occurs through aerosolization of virus-containing excretions, especially respiratory exudates, and virus may be shed for up to 90 d after infection (Greene and VandeVelde 2012). Subclinical infections may occur in up to 70% of domestic dogs (Greene and VandeVelde 2012). In the acute, generalized form, severe respiratory and gastrointestinal signs may be followed in 1–3 wk by neurologic impairment and a high mortality rate (Greene and VandeVelde 2012). If the animal recovers, immunity is long-lived, probably for life (Greene and VandeVelde 2012). Red wolves (Canis rufus) maintain a titer for 3 yr after vaccination (Anderson et al. 2014). In naïve populations, all ages will be affected in CDV epizootics, whereas, in populations with either natural or vaccine-induced immunity, disease occurs chiefly in juveniles after maternally acquired immunity wanes at 3–6 mo old (Greene and VandeVelde 2012).

Canine parvovirus (CPV), a group of closely related parvoviruses (Parvoviridae), was first detected in clinical cases in domestic dogs in 1978. However, seropositive samples were collected from dogs in Greece in 1974 (Steinel et al. 2001) and wolves in the US in 1973 (Mech et al. 2008), and CPV is thought to have negatively affected the wolf population on Isle Royale, Michigan (Mech et al. 2008). The age of the animal at infection is an important determinant of the pathogenesis and ultimate outcome because parvovirus targets rapidly dividing cell lines for replication (Steinel et al. 2001). Infection in neonatal puppies results in fatal myocarditis (Steinel et al. 2001). In immature animals >3 wk old, infection typically causes hemorrhagic gastroenteritis and lymphopenia, whereas subclinical or mild disease is common in adults (Steinel et al. 2001). A strong, long-lasting immunity to the virus develops upon recovery from infection (Steinel et al. 2001).

Mexican wolves, especially the young of the year, are captured annually to be fit for radiotelemetry collars, to administer vaccinations and parasiticides, and to collect samples for health and genetic testing. Capture, handling, and processing of wolves were performed according to the Mexican Wolf Reintroduction Project standard operating procedures. Wolves were captured from the Mexican Wolf Experimental Population Area in eastern Arizona and western New Mexico (33°53′16″N, 109°10′17″W) by helicopter darting or with padded leghold traps. All wolves were marked at first capture with an implanted, passive, integrated transponder.

Blood was collected from the cephalic or lateral saphenous vein and kept chilled until processing. Samples were centrifuged, and the serum was separated from the clot within 24 h. Sera were submitted to the New Mexico Veterinary Diagnostic Laboratory (Albuquerque, New Mexico, USA). Before July 2012, antibodies to CDV were detected with serum virus neutralization or immunofluorescent antibody detection. After July 2012, an enzyme-linked immunosorbent assay was used for the CDV test. Antibodies to CPV were detected via hemagglutination inhibition in all years. Positive cutoff values were established by the laboratory (CDV titers >1:20 for hemagglutination or 1:16 for enzyme linked immunosorbent assay, and CPV titers >1:80).

We used published demographic data (Siminski 2016) and the annual reports of the reintroduction program (Arizona Game and Fish Department 2017) to identify wolves that were born in the reintroduction area or that were released with their dam at <3 mo old from 2002 to 2014 (n=214). We removed from the data set animals that disappeared for unknown reasons before 2 yr old (n=26). Six animals that survived to 2 yr but then disappeared before reaching 4 yr old were included in calculations of survival to 2 yr old but were excluded when calculating the proportion of animals that died between 2 yr and 4 yr old. Wolves lethally removed under a management decision (n=8), legally shot (n=1), illegally shot (n=34), or permanently returned to captivity before 2 yr old (n=27) were removed from the data set. We evaluated the proportion of the remaining 118 wolves that died in each of three age groups: dying at ≤2 yr old (n = 26), dying between 2 yr and 4 yr old (n=24), and surviving ≥4 yr (n=68). We then calculated the proportion of wolves that died in those age groups by birth year, with mortality in the youngest group covering the period from first capture (typically several months of age) to 2 yr old. Because mortality data was truncated for the 2013–14 cohorts, we calculated only the proportion of mortalities for wolves <2 yr old. Odds ratios were used to determine whether mortality rates differed between age categories and genders.

We used mixed-effects logistic regression to estimate the effect of previous exposure to CDV or CPV on mortality of young wolves by using results of serologic testing conducted on samples collected from unvaccinated wolves <2 yr old (as determined by size, dentition, and pack records) captured from 2003 to 2016 (n=108), which served as a proxy for the presence of the viruses. The response variable was binary and indicated whether each wolf died before reaching 2 yr of age. We used random effects to account for variation in mortality among birth cohorts and variation in mortality among packs. We also used fixed effects to control for other variables likely to affect mortality of wolves. Specifically, we included covariates to control for sex and the age at which the wolf was tested, under the expectation that younger wolves were less likely to survive to 2 yr old. Two approaches were used to estimate the effect of exposure to CDV or CPV. First, for 61 wolves lacking individual serologic results but with serologic data at the pack level and, therefore, likely exposed to CPV or CDV, we used a binary covariate indicating whether CDV or CPV seroprevalence was present in the pack to predict death before age 2 yr. Then, for 55 wolves with individual serologic results, we used a binary covariate indicating whether each wolf had CDV or CPV seroprevalence to predict death before age 2 yr. For those analyses, we excluded animals that were tested after death or after age 2 yr or that lacked data regarding the date of testing. For each approach, we used Akaike information criterion with a correction for small sample sizes to identify the best-supported model for explaining the probability of death before age 2 yr. We conducted analyses in the program R (version 3.2.5, R Foundation for Statistical Computing, Vienna, Austria) using the lme4 package (version 1.1.12; Bates et al. 2015) for mixed-effects models.

Annual seroprevalence rates for animals <2 yr old ranged from 0% to 100% (median, 0%) for CDV and from 33% to 100% (median, 89%) for CPV (Table 1). Fewer than five animals were sampled in four of the 14 yr. All packs and nearly all individuals (82%, 89/108) tested were seropositive to CPV. Serologic response to CDV was not detected in animals <2 yr old from 2007 to 2009 (n=14) or from 2011 to 2014 (n=48), but 63% (5/8) wolves were seropositive to CDV in 2006. The odds of mortality were significantly different for males vs. females 2–4 yr old (odds ratio=0.243, P=0.016) but not for male vs. female <2 yr old (odds ratio=1.43, P=0.443) or between all wolves <2 yr old and 2–4 yr old (odds ratio=1.68, P=0.132). The proportion of wolves dying at <2 yr old and at 2–4 yr old varied by cohort year with a greater proportion dying <4 yr old in 2007, 2008, and 2009 (Fig. 1). Age at testing had a negative effect on mortality, indicating that younger animals were less likely to survive until 2 yr old, whereas sex had little effect on mortality. Mortality rates varied across years but not among packs. The best-supported model excluded any effect of CPV or CDV on death before 2 yr at both the pack and individual level (Tables 2, 3).

Table 1

Numbers of unvaccinated Mexican wolves (Canis lupus baileyi) born in the Mexican Wolf Experimental Population Area in Arizona and New Mexico with positive titers for canine distemper virus (CDV) and canine parvovirus (CPV). Only wolves younger than 2 yr old when sampled are included. No samples were collected in 2004. Before July 2012, antibodies to CDV were detected via serum virus neutralization or immunofluorescent antibody detection. After July 2012, an enzyme-linked immunosorbent assay was used for the CDV test. Antibodies to CPV were detected via hemagglutination inhibition in all years.

Numbers of unvaccinated Mexican wolves (Canis lupus baileyi) born in the Mexican Wolf Experimental Population Area in Arizona and New Mexico with positive titers for canine distemper virus (CDV) and canine parvovirus (CPV). Only wolves younger than 2 yr old when sampled are included. No samples were collected in 2004. Before July 2012, antibodies to CDV were detected via serum virus neutralization or immunofluorescent antibody detection. After July 2012, an enzyme-linked immunosorbent assay was used for the CDV test. Antibodies to CPV were detected via hemagglutination inhibition in all years.
Numbers of unvaccinated Mexican wolves (Canis lupus baileyi) born in the Mexican Wolf Experimental Population Area in Arizona and New Mexico with positive titers for canine distemper virus (CDV) and canine parvovirus (CPV). Only wolves younger than 2 yr old when sampled are included. No samples were collected in 2004. Before July 2012, antibodies to CDV were detected via serum virus neutralization or immunofluorescent antibody detection. After July 2012, an enzyme-linked immunosorbent assay was used for the CDV test. Antibodies to CPV were detected via hemagglutination inhibition in all years.
Figure 1

Proportion of Mexican wolves (Canis lupus baileyi) in the Mexican Wolf Experimental Population Area in Arizona and New Mexico that died before reaching 2 yr of age, that died between 2 yr and 4 yr old, or that survived beyond 4 yr, by birth cohort year. Removed from the data set were wolves that were illegally or legally killed or that were permanently removed from the population before 2 yr old. Canine distemper antibodies were detected via serum virus neutralization or immunofluorescent antibody detection in 2003–06 and 2010. In 2014, one wolf died at just older than 2 yr, indicated by an asterisk.

Figure 1

Proportion of Mexican wolves (Canis lupus baileyi) in the Mexican Wolf Experimental Population Area in Arizona and New Mexico that died before reaching 2 yr of age, that died between 2 yr and 4 yr old, or that survived beyond 4 yr, by birth cohort year. Removed from the data set were wolves that were illegally or legally killed or that were permanently removed from the population before 2 yr old. Canine distemper antibodies were detected via serum virus neutralization or immunofluorescent antibody detection in 2003–06 and 2010. In 2014, one wolf died at just older than 2 yr, indicated by an asterisk.

Table 2

Mixed-effects logistic regression models estimating the effect of the presence of canine distemper virus (CDV) or canine parvovirus (CPV) seroprevalence in each pack of Mexican wolves (Canis lupus baileyi) in the Mexican Wolf Experimental Population Area in Arizona and New Mexico on death before 2 yr of age. Seroprevalence was determined by testing samples from wolves younger than 2 yr and only at the first capture. Before July 2012, antibodies to CDV and were detected via serum virus neutralization or immunofluorescent antibody detection. After July 2012, an enzyme-linked immunosorbent assay was used for the CDV test. Antibodies to CPV were detected via hemagglutination inhibition in all years. Models ranked according to Akaike information criterion with a correction for small sample sizes (AICc). The response variable was binary and indicated whether each wolf died before 2 yr of age. We included cohort as a random effect to account for variation in mortality among birth cohorts. We included covariates for sex and the age at which the wolf was tested under the expectation that younger wolves were less likely to survive to 2 yr old.

Mixed-effects logistic regression models estimating the effect of the presence of canine distemper virus (CDV) or canine parvovirus (CPV) seroprevalence in each pack of Mexican wolves (Canis lupus baileyi) in the Mexican Wolf Experimental Population Area in Arizona and New Mexico on death before 2 yr of age. Seroprevalence was determined by testing samples from wolves younger than 2 yr and only at the first capture. Before July 2012, antibodies to CDV and were detected via serum virus neutralization or immunofluorescent antibody detection. After July 2012, an enzyme-linked immunosorbent assay was used for the CDV test. Antibodies to CPV were detected via hemagglutination inhibition in all years. Models ranked according to Akaike information criterion with a correction for small sample sizes (AICc). The response variable was binary and indicated whether each wolf died before 2 yr of age. We included cohort as a random effect to account for variation in mortality among birth cohorts. We included covariates for sex and the age at which the wolf was tested under the expectation that younger wolves were less likely to survive to 2 yr old.
Mixed-effects logistic regression models estimating the effect of the presence of canine distemper virus (CDV) or canine parvovirus (CPV) seroprevalence in each pack of Mexican wolves (Canis lupus baileyi) in the Mexican Wolf Experimental Population Area in Arizona and New Mexico on death before 2 yr of age. Seroprevalence was determined by testing samples from wolves younger than 2 yr and only at the first capture. Before July 2012, antibodies to CDV and were detected via serum virus neutralization or immunofluorescent antibody detection. After July 2012, an enzyme-linked immunosorbent assay was used for the CDV test. Antibodies to CPV were detected via hemagglutination inhibition in all years. Models ranked according to Akaike information criterion with a correction for small sample sizes (AICc). The response variable was binary and indicated whether each wolf died before 2 yr of age. We included cohort as a random effect to account for variation in mortality among birth cohorts. We included covariates for sex and the age at which the wolf was tested under the expectation that younger wolves were less likely to survive to 2 yr old.
Table 3

Mixed-effects logistic regression models estimating the effect of the presence of canine distemper (CDV) or canine parvovirus (CPV) seroprevalence as determined by testing samples from wolves younger than 2 yr old and only at the first capture in individual Mexican wolves (Canis lupus baileyi) in the Mexican Wolf Experimental Population Area in Arizona and New Mexico on death before 2 yr of age. Before July 2012, antibodies to CDV were detected via serum virus neutralization or immunofluorescent antibody detection. After July 2012, an enzyme-linked immunosorbent assay was used for the CDV test. Antibodies to canine parvovirus were detected via hemagglutination inhibition in all years. Models ranked according to Akaike information criterion with a correction for small sample sizes (AICc). The response variable was binary and indicated whether each wolf died before reaching 2 yr of age. We included cohort as a random effect to account for variation in mortality among birth cohorts. We included covariates for sex and the age at which the wolf was tested under the expectation that younger wolves were less likely to survive to 2 yr old.

Mixed-effects logistic regression models estimating the effect of the presence of canine distemper (CDV) or canine parvovirus (CPV) seroprevalence as determined by testing samples from wolves younger than 2 yr old and only at the first capture in individual Mexican wolves (Canis lupus baileyi) in the Mexican Wolf Experimental Population Area in Arizona and New Mexico on death before 2 yr of age. Before July 2012, antibodies to CDV were detected via serum virus neutralization or immunofluorescent antibody detection. After July 2012, an enzyme-linked immunosorbent assay was used for the CDV test. Antibodies to canine parvovirus were detected via hemagglutination inhibition in all years. Models ranked according to Akaike information criterion with a correction for small sample sizes (AICc). The response variable was binary and indicated whether each wolf died before reaching 2 yr of age. We included cohort as a random effect to account for variation in mortality among birth cohorts. We included covariates for sex and the age at which the wolf was tested under the expectation that younger wolves were less likely to survive to 2 yr old.
Mixed-effects logistic regression models estimating the effect of the presence of canine distemper (CDV) or canine parvovirus (CPV) seroprevalence as determined by testing samples from wolves younger than 2 yr old and only at the first capture in individual Mexican wolves (Canis lupus baileyi) in the Mexican Wolf Experimental Population Area in Arizona and New Mexico on death before 2 yr of age. Before July 2012, antibodies to CDV were detected via serum virus neutralization or immunofluorescent antibody detection. After July 2012, an enzyme-linked immunosorbent assay was used for the CDV test. Antibodies to canine parvovirus were detected via hemagglutination inhibition in all years. Models ranked according to Akaike information criterion with a correction for small sample sizes (AICc). The response variable was binary and indicated whether each wolf died before reaching 2 yr of age. We included cohort as a random effect to account for variation in mortality among birth cohorts. We included covariates for sex and the age at which the wolf was tested under the expectation that younger wolves were less likely to survive to 2 yr old.

North American grey wolf (Canis lupus) populations have high seroprevalences to CPV with little annual variability and lower and more-variable seroprevalences for CDV (Gese et al. 1997; Almberg et al. 2009; Nelson et al. 2012; Watts and Benson 2016). However, Mech et al. (2008) detected a cyclic variation in CPV seroprevalence and an association with a decline in the proportion of pups in the population; CDV seroprevalence was not assessed in that study. Almberg et al. (2009) found that the probability of survival, especially of pups, was significantly lower in years with the highest seroprevalence to CDV, whereas we did not find a relationship.

A number of variables possibly affected our analyses. Management protocols and monitoring levels of the Mexican wolf population have changed, trending for increased monitoring of young. Den monitoring before 2012 occurred by chance or when there was a perceived need. After 2012, except for the dens of primiparous females, a pup count was made between their birth and approximately 3 wk of age. However, because only those animals with studbook numbers, assigned when a wolf is first sampled and marked at a minimum of 6 mo old, were included in the study, that change should not have affected mortality rates from year to year. Similarly, the change in serologic testing methods could have affected apparent seroprevalence rates if the methods differed in sensitivity or specificity (Thrusfield 2005). Because wolf pups that died in the den were not entered into the studbook and, thus, were not included in our data, we may have underestimated the effects of disease on mortality rates because CDV and CPV are more virulent in very young pups (Steinel et al. 2001; Greene and VandeVelde 2012).

The strength of our analysis was limited by the small sample size for some years (Table 1). Although the proportion of wolves dying before reaching 2 yr old varied, the regression analysis did not identify a significant correlation with seroprevalence of CDV, possibly because seroprevalence is an imprecise proxy for the occurrence of disease. Seroprevalence is an inexact measure of disease prevalence because titers persist for several years (Steinel et al. 2001; Greene and VandeVelde 2012; Anderson et al. 2014). If the infectious agent has a high fatality rate, then seroprevalence in the population will be low or absent (Gese et al. 1997). Long-term persistence of titers is thought to occur with CPV and CDV in Canis spp. (Gese et al. 1997; Deem et al. 2000; Steinel et al. 2001; Greene and VandeVelde 2012).

After our study, an increase in seroprevalence to CDV was detected in 2015 (18%, 2/11) and in 2016 (40%, 4/10) in unvaccinated wolves. In January–March 2017, two subadult (10 mo old), male wolves were observed to be emaciated and exhibiting neurologic signs. Both died and, on necropsy, had lesions consistent with CDV and were positive for CDV by PCR.

With seroprevalence used as a proxy for the presence of CDV and CPV in the population, statistical modeling did not identify an effect of either virus on Mexican wolf mortality rates. The primary goal of the mortality investigations was to identify wolves that were killed illegally, and disease testing was not part of the protocol. Our understanding of the effects of these diseases on wolf population dynamics would be improved by including testing for infectious diseases in mortality investigations, continuing serologic testing of juveniles, checking dens shortly after whelping to determine actual litter size, and investigation of pup survival to weaning through remote observation.

We thank Susan Dicks, Peter Alcumbrac, John Oakleaf, and members of the Interagency Field Team and the Mexican wolf reintroduction management group for collecting samples and providing laboratory results. We thank the US Fish and Wildlife Service Forensic Laboratory for determining when wolves had been illegally shot. Funding for this research came from US Fish and Wildlife Service State Wildlife Grant and from the Wildlife Restoration Act Project W-78-R.

LITERATURE CITED

LITERATURE CITED
Almberg
ES
,
Mech
LD
,
Smith
DW
,
Sheldon
JW
,
Crabtree
RL.
2009
.
A serological survey of infectious disease in Yellowstone National Park's canid community.
PLoS One
4
:
e7042
.
Anderson
K
,
Case
A
,
Woodie
K
,
Waddell
W
,
Reed
HH.
2014
.
Duration of immunity in red wolves (Canis rufus) following vaccination with a modified live parvovirus and canine distemper vaccine.
J Zoo Wildl Med
45
:
550
554
.
Arizona Game and Fish Department
.
2017
.
Mexican wolf reintroduction & management.
.
Bates
D
,
Mächler
M
,
Bolker
B
,
Walker
S.
2015
.
Fitting linear mixed-effects models using lme4.
J Stat Softw
67
:
1
48
.
Deem
SL
,
Spelman
LH
,
Yates
RA
,
Montali
RJ.
2000
.
Canine distemper in terrestrial carnivores: A review.
J Zoo Wildl Med
31
:
441
451
.
Gese
EM
,
Schultz
RD
,
Johnson
MR
,
Williams
ES
,
Crabtree
RL
,
Ruff
RL.
1997
.
Serological survey for diseases in free-ranging coyotes (Canis latrans) in Yellowstone National Park, Wyoming.
J Wildl Dis
33
:
47
56
.
Greene
CE
,
VandeVelde
M.
2012
.
Canine distemper.
In
:
Infectious diseases of dogs and cats
, 4th Ed,
Greene
CE
, editor.
Elsevier
,
St. Louis
,
Missouri
, pp.
25
42
.
Mech
LD
,
Goyal
SM
,
Paul
WJ
,
Newton
WE.
2008
.
Demographic effects of canine parvovirus on a freeranging wolf population over 30 years.
J Wildl Dis
44
:
824
836
.
Nelson
B
,
Hebblewhite
M
,
Ezenwa
V
,
Shury
T
,
Merrill
EH
,
Paquet
PC
,
Schmiegelow
F
,
Seip
D
,
Skinner
G
,
Webb
N.
2012
.
Prevalence of antibodies to canine parvovirus and distemper virus in wolves in the Canadian Rocky Mountains.
J Wildl Dis
48
:
68
76
.
Siminski
DP.
2016
.
Mexican wolf, Canis lupus baileyi, international studbook, 2016.
The Living Desert
,
Palm Desert, California
,
115
pp.
Steinel
A
,
Parrish
CR
,
Bloom
ME
,
Truyen
U.
2001
.
Parvovirus infections in wild carnivores.
J Wildl Dis
37
:
594
607
.
Thrusfield
M.
2005
.
Veterinary epidemiology.
3rd Ed.
Blackwell Science
,
Oxford, UK
,
610
pp.
US Fish and Wildlife Service
.
2017
.
Mexican wolf biological report: Version 2, region 2.
US Fish and Wildlife Service
,
Albuquerque, New Mexico
,
226
pp.
Watts
DE
,
Benson
AM.
2016
.
Prevalence of antibodies for selected canine pathogens among wolves (Canis lupus) from the Alaska Peninsula, USA.
J Wildl Dis
52
:
506
515
.
Williams
ES
,
Thorne
ET
,
Appel
MJG
,
Belitsky
DW.
1988
.
Canine distemper in black-footed ferrets (Mustela nigripes) from Wyoming.
J Wildl Dis
24
:
385
398
.