Epidemiology of natural transmission of bovine leukemia virus infection


Material Information

Epidemiology of natural transmission of bovine leukemia virus infection
Bovine leukemia virus infection
Physical Description:
ix, 167 leaves : ill. ; 28 cm.
Thurmond, Mark Cy, 1947-
Publication Date:


Subjects / Keywords:
Bovine leukosis   ( lcsh )
Leukemia in animals   ( lcsh )
Cattle -- Diseases   ( lcsh )
bibliography   ( marcgt )
theses   ( marcgt )
non-fiction   ( marcgt )


Thesis (Ph. D.)--University of Florida, 1982.
Includes bibliographical references (leaves 145-165).
Statement of Responsibility:
by Mark Cy Thurmond.
General Note:
General Note:

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 000316109
notis - ABU2900
oclc - 08555991
System ID:

Full Text







Copyright 1982


Mark Cy Thurmond




Technical assistance with agar-gel immunodiffusion and

animal sampling was provided by Ms. J. Hennemann, Ms. J.

Ring, Mr. C. Maden, Mr. T. O'Donnell, Mr. A. Green, and

Mr. J. Lindsey. Most data management and computer program-

ming were performed by Ms. J. Galvez and Mr. D. Puhr. Edi-

torial assistance was provided by Dr. M. Burridge, Dr. P.

Nicoletti, and Dr. C. Wilcox. The typist was Ms. B.

Smerage. The valuable support and discussions offered by

the following people are gratefully acknowledged: Dr. M.

Burridge, Dr. R. Carter, Dr. M. Drost, Dr. C. Franti

(University of California/Davis), Dr. J. Gaskin, Dr. R.

Kahrs, Dr. J. Miller (USDA, Ames, IA), Dr. K. Portier,

Dr. O. Straub (West Germany), Dr. M. Van Der Maaten (USDA,

Ames, IA), Dr. S. Walter (Yale University), and Dr. C.

Wilcox. Additional thanks is extended to Dr. R. Carter,

Dr. K. Portier, and Dr. C. Wilcox for their interest and

support in statistical designs and analyses. Financial aid

was provided by United States Department of Agriculture

cooperative agreement 58-519B-0-872 and by the Wetterburg

Foundation of Newark, New Jersey.








. 4

Bovine Leukemia Virus Infection .
Bovine Leukemia Virus. .
Pathogenesis . .
Transmission Via Excretions and
Secretions . .
Serology . .
Factors Examined for Associations with
Bovine Leukemia Virus Infection .
Transmission of Bovine Leukemia Virus
Infection . .
Infection in Other Domestic Animals. .
Reviews . .


Population Studied . .
Location and Climate . .
Management Practices . .
Sampling Design . .
Demographic Information .
Serology . .
Other Species Examined .
Diseases or Conditions Observed. .
Data Collection and Computer Programs.


Introduction . .
Materials and Methods .
Results . .
Discussion . .


S 4
S 5
S 6

S 8
S 9

S 9

S. 13
S. 16
S. 17

. viii



Introduction . .. 41
Materials and Methods. . .. 41
Results . .. 44
Discussion . .. 45


Introduction . 53
Materials and Methods . ... 54
Results. . .. 58
Discussion . . 60


Introduction . 83
Materials and Methods. . .. 84
Results. . .. 86
Discussion . ... 87


Introduction . ... 98
Materials and Methods. . .. 99
Results. . . 104
Discussion . . 105


Introduction . .. 118
Materials and Methods. . .. .118
Results. . .. .120
Discussion . .. 120

X SUMMARY . . 129



FLORIDA. . . 135


BETWEEN JULY 1, 1979, AND JUNE 30, 1981,
RESEARCH UNIT. . .. .136


DIFFUSION. . .. .144




Abstract of Dissertation Presented to the Graduate Council
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy



Mark Cy Thurmond

May 1982

Chairman: Michael Burridge
Major Department: Animal Science

A 27-month study examined 473 dairy cattle for associa-

tions between bovine leukemia virus (BLV) infection and host

and environmental factors. Cattle sera were tested at

monthly intervals for BLV antibodies by agar-gel immuno-

diffusion using the glycoprotein-51 antigen. A model of BLV

colostral antibody decay in 130 calves predicted infection

in calves less than six months of age and estimated anti-

body half-life to be 27.1 1.2 days. Colostral antibody

decay did not differ between BLV-infected and noninfected

calves for slope (p= 0.45) or intercept (p= 0.43). By 95

days of age, 50% of the calves had no detectable BLV

colostral antibodies.

Of 125 calves born to BLV-infected cows and followed

for at least four months, eight (6.4%) had precolostral BLV

antibodies, as determined by radioimmunoassay using the

glycoprotein-51 antigen. In utero infection withBLV was


not associated with dam age (p= 0.86), dam parity (p =0.83),

breed (p= 0.66), sex (p= 0.11), or stage of gestation in

which the dam was infected (p= 0.50). Calves infected in

utero did not pose an increased risk of infection to calves

penned next to them (p = 0.61).

Prevalence rates of infection were 9%, 16%, and 63% at

6, 16, and 27 months of age, respectively. Age-specific

rates of infection were not associated with dam age (p =

0.79), dam parity (p= 0.75), dam BLV-status (p= 0.46), breed

(p= 0.86), or BLV-status of colostrum consumed (p= 0.50).

An algorithm was described which allocated probabili-

ties of infection to locations occupied by an animal prior

to detection of infection. Small calf pastures were associ-

ated with less infection than was the calf barn (p < 0.05).

No less infection was associated with individual outdoor

calf pens compared to contiguous indoor pens (p> 0.05).

Risk of infection associated with the dry herd was five

times that for heifer pastures (p< 0.0001) and accounted

24 infections per 100 heifers per 100 days at risk.

Vaccination for infectious diseases was not associ-

ated with increased BLV infection (p= 0.33). Infection

rates were not associated with month of birth (p= 0.24) or

with season of potential arthropod vectors (p= 0.20).

Heifer infection was likely to occur in late winter or

spring (p = 0.01).


Bovine leukemia virus (BLV) has been shown to be the

causative agent of enzootic bovine leukosis, a neoplastic

disease of cattle (Callahan et al., 1976; Kettmann et al.,

1976; Miller et al., 1969). Bovine leukosis is believed to

have spread to western European countries from the Baltic

region during World War I (Bendixen, 1965). Following

World War II, efforts were undertaken in some European

countries to reduce the tumor incidence rate through hemato-

logic examinations for persistent lymphocytosis, a phenomenon

associated with bovine lymphosarcoma (Bendixen, 1965).

After discovery of BLV, and subsequent development and use

of serologic methods for mass screening (Hoff-Jorgensen et

al., 1978; Miller et al., 1969; Miller and Van Der Maaten,

1976a; Onuma et al., 1975; Schmidt et al., 1978), eradica-

tion of enzootic bovine leukosis progressed rapidly (Bause

et al., 1978; Mammerickx et al., 1978a; Straub, 1978b).

In order to preserve gains made in these programs,

restrictions were placed on BLV-seropositive cattle and on

semen entering countries either free from BLV or with BLV

control programs (Miller, 1980). Such restrictions have

placed an economic burden on the cattle export markets of

the United States (Mix, 1979). Because of the high genetic

quality of American cattle, eradication or control of BLV

infection using European methods of test and slaughter

would not be a pragmatic alternative for the American pro-

ducer. Interest, therefore, has focused on prevention of

transmission and test and segregation within a herd (Miller

and Van Der Maaten, 1978a).

A prerequisite to control of BLV in a herd is a clear

understanding of natural transmission of infection from

fetal life to adulthood or to the age at which heifers

would move to export markets. Several constraints make the

study of natural transmission patterns difficult and may

explain the lack of reports of long-term prospective studies

in the literature. A major obstacle is the necessity for a

large sample of animals to be tested at close intervals

over a long time period. At the same time changes in man-

agement or environmental factors must be recorded.

The device used to measure infection must be sensitive,

specific, inexpensive, simple, and meet specifications of

other programs. This necessitates the use of agar-gel im-

munodiffusion because it fulfills the above conditions

(Miller, 1980). Definition of infection by a serologic

test, however, has important limitations. For instance,

discrimination has not been made between colostral anti-

bodies and infection-induced antibodies in calves less

than six to seven months of age which consumed colostrum

from a BLV-infected cow (Ferrer et al., 1977b). Another

problem in defining infection is that seroconversion may

lag behind BLV infection by as much as two to three months

(Mammerickx et al., 1980; Straub, 1978b; Van Der Maaten and

Miller, 1978b, 1978c).

These constraints are not unique to the study of the

epidemiology of BLV transmission. It is important, there-

fore, that designs for the study and eventual control of

BLV be generally applicable to the study of other diseases

and infections.

Examination of risks of BLV infection in a large cattle

population over a 27-month period is presented here as a

logical progression from the fetal environment to adult-

hood. Detection of BLV infection based on serologic cri-

teria is used as a proxy for infection. The intent is not

to attempt statements about specific routes of BLV infec-

tion, but to describe temporal and spatial patterns of

natural infection observed in animals studied. Furthermore,

factors possibly associated with deviations in those pat-

terns will be examined using existing analytic methods in

a framework applicable to other diseases. In addition, new

techniques are presented which improve the efficiency of a

serologically-based diagnosis.


Bovine Leukemia Virus Infection

Clinical Appearance

Manifestations of infection with BLV vary from no

signs to persistent lymphocytosis or to lymphosarcoma

(Abramova et al., 1974; Grimshaw et al., 1979; Kenyon and

Piper, 1977; Kumar et al., 1978; Sorenson, 1979; Stober,

1968). Clinical signs of tumor involvement usually are

seen in cattle over five years of age and are referable to

the organ system involved (Abramova et al., 1974, Grimshaw

et al., 1979; Sorenson, 1979; Stober, 1968).

Sporadic bovine leukosis (i.e., juvenile, thymic, or

cutaneous leukosis) is not associated with BLV infection

(Bundza et al., 1980; Chander et al., 1977; Onuma, 1978;

Onuma et al., 1979; Richards et al., 1981; Straub and

Weiland, 1977).

In the absence of tumor involvement, BLV-infected

animals do not appear to suffer production losses (Langston

et al., 1978).


Bovine leukemia virus infection is a ubiquitous infection

throughout the world (Burny et al., 1980; Burridge et al.,


1981). A survey of cattle in the state of Florida recently

estimated the infection rate of BLV in dairy cattle to be

48% (Burridge et al., 1981).

Bovine Leukemia Virus

Discovery of BLV was made following phytohemagglutinin-

stimulation of lymphocytes from cattle with lymphosarcoma

(Miller et al., 1969). The virus is classified as a single-

stranded RNA retrovirus (Burny et al., 1980). It is spheri-

cal in shape with a diameter of 60-125 nm (Calafat et al.,

1974; Calafat and Ressang, 1977a, 1977b; Dutta et al.,

1970; Miller et al., 1969).

Several viral proteins have been described. There are

at least two glycoproteins, gp-30 (Dietzschold et al.,

1978) and gp-51 (Onuma et al., 1975), which constitute the

outer shell of BLV (Burny et al., 1980; Devare and

Stephenson, 1977; Driscoll et al., 1977). An ether-

resistant protein constitutes the internal or core anti-

gen, known as p-24 (Gilden et al., 1975; McDonald and

Ferrer, 1976; Miller and Olson, 1972). The BLV genome

codes for a reverse transcriptase which has a unique re-
quirement for Mg (Dietzschold et al., 1978; Gilden et

al., 1975; Graves et al., 1977; Kettmann et al., 1976).

Six mutant strains of BLV have been investigated recently

(Couez et al., 1981; Kettmann et al., 1981).

Several investigators have identified BLV as a C-type

virus (Burny et al., 1980; Dutta et al., 1970; Ferrer et al.,

1971; Kawakami et al., 1970; Mussgay et al., 1977; Stock

and Ferrer, 1972; Van Der Maaten et al., 1974; Weiland

and Ueberschar, 1976). Others have been reluctant to de-

scribe BLV as a B- or C-type virus since immature viruses

are rarely found outside the cell (Calafat et al., 1974;

Calafat and Ressang, 1977a, 1977b; Dekegel et al., 1977).

Comparisons of BLV with other retroviruses or onco-

viruses by molecular hybridization have shown that BLV is

biochemically distinct from Friend mouse leukemia virus and

visna maedi virus (Kaaden et al., 1977), Rauscher leukemia

virus (Kettmann et al., 1975, 1976), simian sarcoma (wooly

monkey) virus, murine sarcoma virus, feline sarcoma virus,

and feline leukemia virus (Kettmann et al., 1975, 1977).

Other studies have demonstrated a lack of cross-reactivity

between proteins of BLV and Mason Pfizer monkey virus

(McDonald and Ferrer, 1976; McDonald et al, 1976), and

bovine syncytia virus (McDonald et al., 1976).


Tissues Involved

Replication of BLV occurs mainly in B-lymphocytes

(Kenyon and Piper, 1977; Paul et al., 1977), but an associa-

tion with T-lymphocytes also has been reported (Takashima

et al., 1977). Further support for B-cell involvement was

found in the expansion of the B-cell population in BLV-

infected cattle (Kenyon and Piper, 1977; Kumar et al.,


Following subcutaneous inoculation of leukocytes from

a BLV-infected steer, BLV was isolated from the spleen

after eight days, from leukocytes after 14 days, and oc-

casionally from prescapular lymph nodes thereafter (Van Der

Maaten and Miller, 1978b). In that study, BLV could be

isolated from lymphocytes two to three weeks before a de-

tectable serologic response, and virus was not isolated

from the thymus. As few as 2500 washed lymphocytes from an

infected steer have been able to transmit BLV infection to

susceptible calves (Van Der Maaten and Miller, 1978c).

Integration of BLV in the Host Cell Genome

Results of studies using BLV-specific DNA probes sug-

gest that BLV is an exogenous virus (Callahan et al., 1976;

Deschamps et al., 1981; Kettmann et al., 1976, 1978a, 1978b,

1979a; Kukaine et al., 1979). The DNA from lymphocytes of

BLV-infected cattle has viral sequences that cannot be

identified in the DNA from lymphocytes of noninfected cat-

tle (Callahan et al., 1976; Kettman et al., 1976), or in

normal cell DNA from BLV-infected cows (Kettmann et al.,


The BLV provirus is integrated in several sites of

the lymphocyte DNA in cattle with persistent lymphocytosis,

but in only one or a few sites in the DNA of cells of

lymph node tumors (Kettmann et al., 1979a, 1980a, 1980b).

Less than 5% of peripheral lymphocytes in asymptomatic,

infected cattle contain the provirus, whereas up to 33% of

the circulating lymphocytes in cattle with persistent

lymphocytosis contain the provirus (Kettmann et al., 1980b).

Seroconversion Period

Seroconversion following infection with BLV has been

found to occur between two and seven weeks in cattle ex-

perimentally inoculated (Mammerickx et al., 1980; Van Der

Maaten and Miller, 1978a, 1978b). Half of the animals in

these studies had seroconverted by five weeks postinocula-

tion. The seroconversion period was similar for sheep

experimentally inoculated by the intradermal route, oral

route, or by BLV-carrying tabanid flies (Gentile and Rutili,

1978; Mammerickx et al., 1980; Ohshima et al., 1981). The

seroconversion period for animals naturally infected is

considered to be less than three months (Straub, 1978b),

and the pattern of seroconversion is believed to be similar

to that for experimentally infected animals (Van Der Maaten

and Miller, 1978b).

Transmission Via Excretions and Secretions

It is well documented that BLV can be experimentally

transmitted to cattle via blood (Mammerickx et al., 1980;

Van Der Maaten et al., 1981a) and lymphocytes (Miller and

Van Der Maaten, 1978b; Van Der Maaten and Miller, 1978a,

1978b) from infected animals. The virus has been demon-

strated in saliva but not in prostatic fluid or feces of

infected cattle (Ressang et al., 1980). The p-24 antigen

of BLV has been found'in urine of naturally infected

animals (Gupta and Ferrer, 1980).

Semen collected by manual massage from a BLV-infected

bull transmitted BLV infection to susceptible sheep (Lucas

et al., 1980). Another study failed to demonstrate BLV in

semen collected from BLV-infected bulls following normal

ejaculation (Miller and Van Der Maaten, 1979).


Several serologic tests for the detection of BLV have

been described for both the gp-51 and the p-24 antigen

(Burny et al., 1980). Agar-gel immunodiffusion using gp-51

has been recommended for use by member countries of the

European Economic Community (Kaaden and Stephenson, 1978),

because of its high sensitivity, simplicity, and low cost

(Miller, 1980). Recently a radioimmunoassay procedure

was described using gp-51 (Bex et al., 1979). This test

may be the most sensitive one presently available (Miller

et al., 1981).

Factors Examined for Associations with Bovine
Leukemia Virus Infection

Genetic Susceptibility

A genetic predisposition to enzootic bovine leukosis

and bovine lymphosarcoma was suspected before discovery of

BLV. Leukosis was observed more frequently in daughters

of affected cows than in daughters of unaffected cows

(Bendixen, 1965; Larson et al., 1970). Pedigree studies of

lymphosarcoma found clustering of cases by sire and/or dam

families (Crowshaw et al., 1963; Cypess et al., 1974;

Marshak et al., 1962). It also was observed that herds

which were inbred experienced higher rates of leukosis and

lymphosarcoma than did noninbred herds (Abt, 1968; Laktionov

and Nakhmanson, 1972), but purebred herds were found to

have lower rates of leukosis than nonpurebred herds

(Anderson et al., 1971).

One study estimated the heritability of susceptibility

to BLV infection to be 0.44 0.22 (Burridge et al., 1979).

A study of lymphosarcoma, however, was not able to associate

the disease with serologically defined antigens controlled

by the BoLA-A locus (Takashima and Olson, 1978).

Parental Infection with Bovine Leukemia Virus

The effect of BLV infection of the dam on subsequent

BLV infection in the offspring has been examined by several

investigators. In one report, dam status appeared to have

a significant influence on progeny infection (Baumgartener

et al., 1978), but the authors felt that such an effect may

have been due to high prevalence rates in some herds. In a

longitudinal study, no association was found between dam

status and age at which progeny became infected (Olson et

al., 1978). Reports on cross-sectional studies concluded

that presence of BLV antibodies in the dam was not associ-

ated with subsequent progeny infection (Hofirek, 1980;

Valikhov, 1978).

In a large study of progeny from BLV-infected, AI bulls,

offspring from infected sires did not have as high a rate of

subsequent BLV infection as did those from noninfected sires

(Baumgartener et al., 1978).


Few studies have examined for associations between

BLV infection and sex. Reports in which sex effects were

studied suggested no difference in infection rates between

males and females (Baumgartener et al., 1975; Evermann et

al., 1980).


It has been reported that many different breeds are

susceptible to BLV infection (Burridge et al., 1981; Marin

et al., 1978). Analysis of data from a survey of Florida

cattle suggested that Jerseys had a higher infection rate

than Holsteins (Burridge et al., 1981). However, a study

within a Florida dairy herd indicated no difference existed

between rates of infections for Jerseys and Holsteins

(Burridge et al., 1979).


Age-specific prevalence rates of BLV infection have

been shown to follow a characteristic sigmoidal curve.

Rates increased linearly from one to four years of age,

after which they plateaued (Burridge et al., 1979, 1981;

Chander et al., 1978; Evermann et al., 1980; Ferrer et al.,

1976; Hofirek, 1980; Mammerickx et al., 1978a, 1978b;

Marin et al., 1978; Olson et al., 1973; Piper et al., 1979).

Peak prevalence rates of infection were observed at four

years of age in beef cattle and at more than nine years of

age in dairy cattle (Burridge et al., 1981). Ages at which

a sharp, linear increase in rates occurred varied from study

to study and from herd to herd. In some herds rates of

infection began to level off at two to four years of age

(Burridge et al., 1979; Olson et al., 1973), while in other

herds rates reached a plateau at four to five years of age

(Chander et al., 1978; Hofirek, 1980; Mammerickx et al.,

1978a, 1978b; Marin et al., 1978).

A few studies have approached age-specific rates of

infection in a longitudinal design using birth cohorts.

Results of one of these investigations showed that animals

of similar ages experienced different rates of infection,

according to the birth cohort (Wilesmith et al., 1980).

In the other study, each new 12-month cohort entered the

herd with a lower prevalence rate than did the previous

cohort. Rates within a cohort did not appear to change as

cattle aged. It was suggested further that higher preval-

ence rates of infection observed in older animals in

cross-sectional studies represented high-rate cohorts

(Huber et al., 1981).

Transmission of Bovine Leukemia Virus Infection

In Utero

Rates of natural in utero infection with BLV have been

reported to range from 3% to 25% (Ferrer et al., 1976,

1977a, 1977b; Piper et al., 1979). Stage of gestation dur-

ing which a dam is experimentally infected has not been

associated with the frequency of infection in progeny (Van

Der Maaten et al., 1981b).

Physical Contact

Close physical contact between infected and susceptible

cattle is believed to be a prerequisite to BLV transmission

(Ferrer and Piper, 1981; Maas-Inderwiesen et al., 1978;

Miller and Van Der Maaten, 1978a; Wilesmith et al., 1980).

Newborn calves were more likely to develop leukosis when

placed in close contact with leukotic cows (Straub, 1971).

Infection rates increased during winter months in one herd

studied, suggesting transmission associated with indoor

housing conditions (Wilesmith et al., 1980). Limiting

physical contact by vacating a stall between animals or by

placing a single fence between animals appeared to retard

transmission of infection (Miller and Van Der Maaten, 1978a).

Arthropod Vectors

Bovine leukemia virus has been isolated from the mid-

gut of horseflies after feeding on a BLV-infected cow

(Bech-Nielsen et al., 1978). In an experimental study,

horsefly transmission of BLV infection to sheep was demon-

strated (Ohshima et al., 1981). It has been suggested

that high rates of infection observed in animals during

summer months support the hypothesis of vector-borne trans-

mission of BLV infection (Bech-Nielsen et al., 1978; Onuma

et al., 1980). Another study, however, observed higher

rates of infection during winter months (Wilesmith et al.,


Ixodes ricinus ticks have been suggested as a possible

explanation for geographic differences in rates of infec-

tion in Sweden (Hugoson and Brattstrom, 1980).


Intranasal instillation of BLV-infected lymphocytes

produced infection in one of two calves and an aerosol

exposure to BLV-culture fluids produced infection in two of

two calves (Van Der Maaten and Miller, 1978c). Both methods,

however, also exposed the oral cavity and the latter method

exposed the conjunctivae.


Bovine leukemia virus or BLV-like particles have been

identified in milk and colostrum of BLV-infected cows and

cows with lymphosarcoma (Dutcher et al., 1964; Jensen and

Schidlovsky, 1964; Miller and Van Der Maaten, 1979;

Schulze et al., 1966). Transmission of BLV by the oral

route has been demonstrated by inoculation of leukemic

blood into neonatal lambs (Mammerickx et al., 1976a), by

inoculation of lymphocyte cultures into colostrum-deprived

calves (Miller et al., 1972; Van Der Maaten and Miller,

1978c), and by feeding BLV-infected lymphocytes in colostrum

free of BLV antibodies (Van Der Maaten et al., 1981a).

Oral transmission has not been shown in old lambs (Hoss

and Olson, 1974) or in old calves (Van per Maaten and

Miller, 1978c). Infection did not occur when calves were

fed BLV-infected lymphocytes in colostrum containing BLV

antibodies (Van Der Maaten et al., 1981a). Some studies

concluded that considerable emphasis should be placed on

oral transmission, especially when bloody colostrum or

milk is fed (Parfanovich et al., 1978; Seger and Morgan,

1977). Results of other studies suggest that milk or

colostrum does not constitute a major vehicle for trans-

mission of BLV infection (Ferrer et al., 1976; Ferrer and

Piper, 1978, 1981; Piper et al., 1975, 1979). Pasteuriza-

tion of milk or colostrum for 30 seconds at 600C would

probably inactivate the virus (Baumgartener et al., 1976).


One study demonstrated the venereal transmission of

BLV infection after placing BLV-infected lymphocytes in the

cervical canal of susceptible cows (Van Der Maaten and

Miller, 1978c).


Intradermal inoculation of BLV has been shown to be a

viable means of transmission (Van Der Maaten and Miller,

1978c). Iatrogenic transmission of BLV following blood

sampling procedures has been suspected (Bause et al.,

1978; Maas-Inderwiesen et al., 1978; Wilesmith, 1979).

Premunization of cattle for babesiosis and other vaccina-

tion procedures also have been incriminated in BLV trans-

mission (Hugoson and Brattstom, 1980; Hugoson et al.,

1968; Marin et al., 1978; Stamatovic and Jonavic, 1968).

However, transmission of BLV to sheep following intradermal

Tb testing has not been demonstrated (Roberts et al., 1981).

Infection in Other Domestic Animals

The possibility for BLV infection in species other

than the bovine was first suggested when bovine leukosis

and later bovine lymphosarcoma were transmitted to sheep

(Olson et al., 1972; Wittmann and Urbaneck, 1969). Ex-

perimental infection with BLV has been confirmed in sheep

(Bansal and Singh, 1980; Bex et al., 1979; Gentile and

Rutili, 1978; Hoss and Olson, 1974; Mammerickx et al.,

1976a, 1976b, 1980, 1981; Olson and Baumgartener, 1978;

Van Der Maaten and Miller, 1976a), in goats (Hoss and

Olson, 1974; Mammerickx et al., 1981; Olson et al., 1981),

and in pigs (Mammerickx et al., 1981). Evidence exists

for experimental BLV infection in rabbits, but not in rats

(Bansal and Singh, 1980), and in guinea pigs (Lussier and

Pavilanis, 1969). Serologic evidence for BLV infection in

chimpanzees has been found (Van Der Maaten and Miller,

1976b). Passage of BLV through cattle, sheep, or goats

did not reduce the infectivity of the virus for cattle,

sheep, goats, or pigs (Mammerickx et al., 1981). Sheep

have not been found to shed BLV and thus may not be a

reservoir (Mammerickx et al., 1976b; Van Der Maaten and

Miller, 1976a).


Bovine leukosis, bovine lymphosarcoma, and bovine

leukemia virus infection have been extensively reviewed

(Bendixen, 1965; Burny et al., 1978, 1980; Ferrer, 1977,

1979, 1980a, 1980b; Ferrer et al., 1978, 1979; Hoff-

Jorgensen, 1977; House et al., 1975; Markson, 1979;

Mussgay and Kaaden, 1978; Olson, 1974; Olson et al.,

1970; Reed, 1981; Ruppanner and Paul, 1980; Tyler, 1978;

Van Der Maaten and Miller, 1975).


Population Studied

Cattle studied were those in the University of Florida

Dairy Research Unit (DRU) herd. This population was chosen

because it had been characterized previously by a BLV

prevalence rate of 75% (Burridge et al., 1979). Calves

born from July 1, 1979, through June 30, 1981, were fol-

lowed until death, sale, parturition, or September 30, 1981,

whichever occurred first.

Location and Climate

The DRU was located in north-central Florida (latitude

82030' west, longitude 29040' north). It occupied about 450

hectares of reclaimed pine flatwoods 20 km. northwest of the

University of Florida at Gainesville, Florida (APPENDIX A).

The climate was characterized by hot, humid summers and

cool, dry winters (APPENDIX B).

Management Practices

The DRU was a closed herd used as a research facility.

Two 18-month-old heifers, a Jersey (Ml01) and a Holstein

(M121), entered the herd, however, from a commercial dairy

(Bassett's Dairy, Montecello, FL) on May 5, 1980. Both were

negative for BLV antibodies when they entered the herd. The

Jersey had BLV antibodies by April 25, 1981, and the Hol-

stein remained serologically negative through September 26,


Approximately 200 purebred cows were milked, 1/3 of

which were Jerseys and 2/3 Holsteins. Cattle were allowed

to.graze, but it was necessary to supplement rations with

corn silage; corn, sorghum, ryegrass, or alfalfa greenchop;

and concentrates.

Cattle were bred artificially using commercially avail-

able frozen semen. Some semen was used, however, from

young Jersey control sires raised at the DRU. Calving

season ranged from June through December (APPENDIX C).

Calves were born in small pastures (APPENDIX D, Fig. 1) and

remained with their dams for about 12 hours. They then

were placed in either contiguous, wire pens in the calf

barn (APPENDIX D, Fig. 2) or in individual outdoor pens

(APPENDIX D, Fig. 3). Once in pens, they were fasted for

12 to 24 hours, after which time they were fed either

pooled colostrum (obtained from cows one through three days

postpartum) or bulk-tank milk for the next two to three

days. Prior to weaning at about one month of age, calves

were fed bulk-tank milk once a day. Holsteins received

9 lbs. per day and Jerseys 7 lbs. per day. Calves also

were fed concentrates ad libitum from two or three days of


Electric dehorning, ear tagging, and tattooing were

performed during the first month of life. This usually was

not done in a serial fashion from calf to calf at one ses-

sion, but individually, one or two calves per day.

All heifers were vaccinated for brucellosis,a clostridi-
b c
al diseases, and leptospirosis at four months of age

and for infectious bovine rhinotracheitis (IBR),d bovine

virus diarrhea (BVD),d and parainfluenza-3 (PI-3)d at 11

months of age. Brucellosis vaccination was administered

subcutaneously in the neck and other vaccinations intra-

muscularly in the gluteal area. Cattle were vaccinated in

groups using one or two common 18 gauge 1-1/2 inch needles

and a multidose syringe. The order in which cattle were

vaccinated was not recorded.

At the time of brucellosis vaccination, calves were

wormed eitherwith thiabendazolee or levamisole. The

specific type administered to each calf was not recorded.

aStrain 19, Colorado Serum Co., 4950 New York St., Denver, CO.

Siteguard, Clostridium chauvoei, septicum, haemolyticum,
novyii, sordellii, perfringens Type C and D. Jensen-
Salsbery, Kansas City, MO.
Novalep GHP, Jensen-Salsbery, Kansas City, MO.

RESBO 3, Norden Laboratories Inc., Lincoln, NE.

eOmnizole (paste), Merck Animal Health Division, Merck
& Co., Inc., Rahway, NJ.

Ripercol L, American Cyanamid Co., Princeton, NJ, or
Levasole, Pitman-Moore, Inc., Washington Crossing, NJ.

At two to three months of age, calves were moved from

indoor pens in the calf barn, or from nearby outdoor pens,

to paddocks (APPENDIXD, Fig. 4) or to small pastures (0.3

to 0.6 ha.) (APPENDIX D, Fig. 5). At about 10 months of

age, they were moved to larger pastures (2 to 3 ha.)

(APPENDIX D, Figs. 6 and 7). Cattle in each location were

of similar ages.

Heifers were bred between 13 and 16 months of age,

except during summer months when estrous behavior was less

apparent. At 40-45 days gestation, heifers were placed, for

the first time, with adult cattle in the dry herd. For

heifers under study, this began in early January, 1981.

Holstein heifers were removed from the dry herd on May 1,

1981, and placed in pastures holding only bred Holstein

heifers. Jersey heifers continued to be placed in the dry

herd and remained there until shortly before calving.

Most bull calves were sold a few days after birth, al-

though some were retained for veal studies or semen collec-

tion. Management of heifers was similar to that practiced

on most commercial dairies. Heifers usually were not used

in experimental studies until after freshening.

Fly control was practiced sporadically two or three

times during summer months. Lactating cows were hand-

dusted with a commercial insecticide,a and fly baitb was

aMarlate 50, E. I. DuPont de Nemoud and Co., Inc.,
Wilmington, DE.
Golden Malrin, Starbar Division of Zoecon Corp., 12200
Denton Dr., Dallas, TX.

scattered in the calf barn a few times during peak calf

density in the fall. Pastured heifers were not treated

for flies. Most flies observed in the calf barn were

Musca domestic feeding on decaying grain and soured milk.

Pastured cattle were bothered by large numbers of Haematobia

irritans and Stomoxys calcitrans and some tabanidae. No

ticks or lice were observed during the 27-month study.

Sampling Design

Precolostral blood samples were drawn on nearly all

calves born alive between July 1, 1979, and September 30,

1980. To increase the chance of obtaining precolostral

samples, udder bags were placed on cows about one week

prior to calving (APPENDIX D, Fig. 1). This prevented the

calf from nursing until a blood sample could be drawn. In

order to reduce the period of colostrum deprivation, re-

search field personnel were notified by telephone as soon

as a cow calved. This procedure was instituted in December

1979, and calves born after this time were fed 1-2 liters

of colostrum from their own dam within one hour following

birth. After feeding, the navel was washed with a solution

of 3% Lugol's iodine. Through December 1979, calves were

bled precolostrally, at one week of age, and then at three-

month intervals. After December 1979, calves were bled

precolostrally, at two to three days of age, and then at

aNumbers C4591N-C4593N, Nasco, Fort Atkinson, WI.

monthly intervals beginning some time during the first month

of age. Calves born between September 30, 1980, and June 30,

1981, were not bled precolostrally, but were bled at monthly

intervals beginning sometime during the first month of age.

Calves born after June 30, 1981, were not included in the

study. Cows were bled at two, five, and eight months of

gestation, at parturition, and at one month postpartum.

All blood samples were collected from either the jugular or

intercoccygeal vein using a sterile 20 gauge 1-1/2 inch

needle and Vacutainer tube.b

Demographic Information

The following information was obtained for each calf:

BLV status of dam, age of dam, parity of dam, breed, sex,

birth date, sale date, death date, or freshening date.

Locations and inclusive dates of occupation in pens and

pastures were recorded for all cattle as long as they were

followed. As soon as heifers began entering the dry herd,

all cattle in that group were sampled monthly to determine

the BLV status of animals exposed to heifers being studied.

No interventions were made in routine management of calves

or heifers except for colostrum feeding, navel washing, and

monthly mustering and bleeding.

aNo. 5746, Becton-Dickinson and Co., Rutherford, NJ.

bNo. 6512, Becton-Dickinson and Co., Rutherford, NJ.


Serum samples were tested for presence of antibodies

to BLV glycoprotein-51 antigen (gp-51)a using agar-gel im-

munodiffusion (AGID), as previously described (Burridge

et al., 1979). Gel plates were incubated at 24-27 C in an

humidified chamber for 48 hours before being read with the
aid of a direct light beam. Formation of precipitation

lines of identity with positive control sera indicated

presence of BLV antibodies (APPENDIX E). Serial two-fold

dilutions using phosphate buffered saline c were made of

all positive sera and then these dilutions were retested to

determine end-point titers.

All precolostral sera from calves lost to follow-up

before four months of age and negative on AGID were re-

tested using a radioimmunoassay (RIA) procedure with gp-51

(Bex et al., 1979).d Samples precipitating more than 15%

of labeled antigen were considered positive for BLV anti-

bodies. Persistence of antibodies or an increase in BLV

antibody titer was considered evidence of BLV infection.

Precolostral sera from calves with evidence of BLV infec-

tion at some later date were retested by RIA.

aAntigen supplied in the Leukassay-B kit, Pitman-Moore,
Inc., Washington Crossing, NJ, and by Dr. J. Miller, USDA,
Ames, IA.
bModel 651, American Optical Corp., Buffalo, NY.

cDulbecco's formula, Flow Laboratories, Inc., McLean, VA.

Performed by Dr. M. J. Schmerr, USDA, Ames, IA.

Other Species Examined

The DRU maintained between six and eight sheep and

goats during the 27-month study. These animals were kept

in a small pasture, separated from any cattle under study

by at least two fences. Tests for presence of antibodies

to BLV were negative for all sheep and goats in the fall of

1980 and 1981.

Several cats also inhabited the DRU during the study

period. Three of the tamer cats were bled in October 1981,

and no detectable BLV antibodies were found.

Diseases or Conditions Observed

During the 27-month study, lymphosarcoma was diagnosed

in two cows following postmortem examination. Both were

Jerseys, one six years old and the other five years old.

Although no health records were kept, nearly all neo-

natal calves experienced at least one episode of diarrhea

and bronchopneumonia, and several died. According to clini-

cal and pathological reports, a common cause of death,

particularly prior to December 1979, was septicemia result-

ing from omphalophlebitis and/or hypogammaglobulinemia.

Chronic dysentery was prevalent among calves two to six

months of age. This was probably due, at least in part, to

coccidia and trichostrongyles, as suggested by oocysts and

eggs found on fecal flotation. Many calves in this age

group, particularly bull calves in nutrition trials, were

quite cachectic. Heifers older than six months appeared

less affected.

Data Collection and Computer Programs

Data were transcribed onto computer sheets from which

IBM cards were punched. Files were maintained on tape at

the Northeast Regional Data Center (NERDC), University of

Florida, Gainesville, Florida. Computer systems available

through NERDC for statistical analysis were the Statistical

Analysis System (SAS) version 79.5,a Biomedical Computer

Programs (BMDP), and the McGill University System for

Interacting Computing (MUSIC) version 4.1, 1978.

aSAS User's Guide, SAS Institute Inc., Raleigh, NC.

bBMDP, P-series, University of California Press, Berkeley,



Since BLV is currently considered an exogenous virus

(Callahan et al., 1976; Kettmann et al., 1976, 1978a,

1978b, 1979a, 1979b; Kukaine et al., 1979) and there is

no evidence for natural transmission of BLV via semen

(Baumgartener et al., 1978; Miller and Van Der Maaten,

1979; Ressang et al., 1980), prenatal infection is most

likely a consequence of in utero transmission. Suggested

rates for in utero transmission of BLV infection have ranged

from 3% to 25% (Ferrer et al., 1976, 1977a, 1977b; Piper

et al., 1979). Highest rates were observed, however, in

an inbred herd in which selection for bovine leukemia had

been practiced (Piper et al., 1979). The BLV status of the

dams was not stated in some reports (Ferrer et al., 1976,

1977a, 1977b); therefore, rates could have been higher

had they been stated correctly in terms of calves born to

infected cows.

It is not known what factors, if any, predispose a

fetus to infection with BLV. Stage of gestation in which

the dam was infected has not been associated with fetal

infection (Van Der Maaten et al., 1981b). If predisposing

factors are present, calves and/or dams with the related

characteristic could be segregated, thus reducing post-

natal transmission. Otherwise, efforts could be focused on

other aspects of transmission control.

The purpose of this chapter is to describe observed

rates of natural precolostral antibodies to BLV in a large

sample of random-bred calves and to examine for associa-

tions between those rates and certain characteristics of

the calves and their dams.

Materials and Methods

Data Analyzed

Calves examined were those born to BLV-infected dams

and bled precolostrally. A BLV-infected dam was defined as

one which had BLV antibodies up to one month postpartum.

Calves were divided into two groups: (1) those post-

colostrally positive for BLV antibodies and followed for

four months or more, or postcolostrally negative for BLV

antibodies and followed for one month or more, and (2) those

with postcolostral BLV antibodies and not followed for four

months. A positive precolostral sample was considered

invalid, if, after following the calf for four months or

more, there was no evidence for BLV infection. Evidence

for BLV infection was based on a rising antibody titer

during the first six months of age, on observation of titers

outside the 95% prediction level of decay of BLV colostral

antibodies (CHAPTER V), or on persistence of BLV anti-

bodies beyond six months of age.


Rates of precolostral antibodies in both groups of

calves were examined for associations with breed and sex of

calf. Analyses were performed using the FUNCAT program, a

procedure for analysis of categorical data (Grizzle et al.,

1969) offered by SAS. The model was presence (or absence)

of precolostral BLV antibodies = breed + sex + breed x sex.

The response function was the difference between the pro-

portion of calves with and the proportion without pre-

colostral antibodies to BLV, or, as indicated by SAS, 1 -1.

Differences in age and parity distributions for dams

of calves with precolostral antibodies and dams of calves

without precolostral antibodies were examined using the

NPAR1WAY program of SAS, a two-sample Wilcoxon Rank Sums

test for nonnormal distributions. Fisher's exact test,

from the STATPAK subsystem of MUSIC, was used to calculate

the probability of association between the stage of gesta-

tion of seroconversion to BLV by the dam and occurrence of

precolostral antibodies in calves followed four months or



Of 346 calves born during the period July 1, 1979,

through September 30, 1980, 280 calves were bled

precolostrally and, of those, 223 were from BLV-infected

dams. Because many calves were sold or were not followed

for more than a few months, only 125 calves were born from

BLV-positive cows, bled precolostrally, and followed for

four months or more (Table IV-1). Eight (6.4%) of these

calves had precolostral antibodies and antibodies persist-

ing beyond six months of age (Table IV-1, IV-2). In the

group of 223 calves, 18 (8.1%) had detectable precolostral

antibodies (Table IV-1).

Precolostral antibodies were detected with similar

frequency in Jersey and Holstein calves, both in the fol-

lowed group (p= 0.66) and for all calves (p= 0.80) (Table

IV-3). When all calves were examined, males showed a

significantly higher rate of precolostral antibodies than

did heifers (p=0.04). There was no such association in

the 125 calves followed (p= 0.11) (Table IV-3). No sig-

nificant interactions between breed and sex were detected

for either the group followed (p=0.91) or for all calves

(p =0.64) (Table IV-3).

No statistical differences were detected between dis-

tributions of age (p= 0.86) or parity (p=0.83) for dams

having calves with precolostral antibodies and for dams

having calves without detectable precolostral antibodies

(Table IV-4). Complete serologic records were available

for 76 cows seroconverting before the second month of ges-

tation and for 14 cows seroconverting after the second

month. Those seroconverting after the first trimester of

gestation showed no more tendency to have a calf with pre-

colostral antibodies to BLV than did cows seroconverting

before the first trimester (exact p= 0.50) (Table IV-5).


The ruminant fetus may be particularly susceptible to

infection because (1) syndesmochorial placentation does not

allow transfer of maternal antibodies to the fetus; (2) the

fetal immune system is not fully functional; and (3) micro-

bial activity can take place in fetal cells (Osburn, 1981).

Since placentation restricts passage of antibodies to the

fetus, presence of fetal immunoglobulins is indicative of

infection by, or exposure to, an antigen or of placental

leakage of globulins (Brambell, 1970; Husband et al., 1972).

Although no estimates are available for frequency of

placental leaks, such events are believed to be rare

(Brambell, 1970). This has been confirmed by several

studies of in utero infection based on fetal infection and

fetal or precolostral serology (Braun et al., 1973; Dunne

et al., 1973; Fennestad and Borg-Petersen, 1962; Gibson

and Zemjanis, 1973; Horner et al., 1973; Kniazeff et al.,

1967; Osburn and Hoskins, 1971; Osburn et al., 1974;

Van Der Maaten et al., 1981b).

The main difficulty when studying in utero infection

as it occurs naturally in a population, using serologic

detection, is one of logistics. Efforts should be made to

sample the calf as soon after birth as possible and before

it nurses. For this reason, cows were fitted with udder

bags a week or two before their due-date. However, prior

to December 1979, research personnel were lax about udder

bag procedures and precolostral sampling. As a result,

some calves could have sucked BLV-positive cows without

udder bags. Precolostral samples could be partially veri-

fied, however, by following calves to determine if they

became serologically negative. For purposes of analysis,

two groups of calves were described based on confidence in

precolostral samples. A follow-up time of four months was

selected because 80% of calves consuming colostrum contain-

ing BLV antibodies were negative by 120 days of age (CHAPTER

V),and prediction of infected calves, therefore, could be

made with confidence.

Radioimmunoassay using gp-51 antigen has been shown

to be a highly sensitive test of BLV antibodies and may

detect infected animals up to 10 days sooner than AGID

(Miller et al., 1981). After screening with AGID, negative

sera were retested by RIA to identify calves infected in

late gestation. Additional calves identified as precolos-

trally seropositive by RIA may have consumed colostrum, as

previously mentioned. More confidence, therefore, should

be placed on the precolostral antibody rate of 6.4% than on

that of 8.1% because of these logistical problems.

Since reported in utero infection rates of 10% or

more may have counted calves born to negative dams, those

figures represent minimal rates for calves born to infected

dams. In addition, syncytium induction assay, virus neu-

tralization, and/or immunofluorescence assay were used in

those studies to detect BLV or BLV antibodies. These tests

may be less sensitive than RIA with gp-51 for detection of

BLV infection (Miller et al., 1981; Van Der Maaten and

Miller, 1977), and, therefore, in utero infection may have

been underestimated further.

The most likely explanation for the discrepancy between

rates observed in this study and those reported in other

studies rests in innate features of the populations ex-

amined. Higher rates (e.g., 14%, 18%, and 25%) were ob-

served in a Jersey herd in which animals had been inbred

and selected for leukemia (Piper et al., 1979). Such high

rates may suggest a weakening of placental integrity due to

inbreeding allowing for passage of maternal blood into the

fetus. In any case, it is unlikely that the higher rates

are representative of in utero infection in a typical dairy


A rate of 3% was observed in 37 calves from a multiple-

case herd (Ferrer et al., 1976). Again the BLV status of

dams was not reported, and, therefore, rates in calves from

infected cows were possibly higher. Furthermore, caution

should be taken in considering a rate based on only one in-

fected calf from a small sample of 37 calves.

Results of this study indicated that the risk of BLV

transmission did not differ between older, multiparous cows

and younger or primiparous cows. This suggests that

transmission is not a consequence of loss of placental

integrity associated with increasing physiologic age or

pregnancies. In addition, stage of gestation at maternal

infection did not appear to influence fetal infection rates,

in accordance with observations from experimental studies

(Van Der Maaten et al., 1981b).

Reports on studies of breed effects on BLV infection

rates indicate that many breeds are susceptible (Burridge

et al., 1981; Marin et al., 1978). One report concluded

that higher infection rates occurred in Jerseys (Burridge

et al., 1981), but these differences could have been at-

tributable to high rates in particular age cohorts (Huber

et al., 1981; Wilesmith et al., 1980). Furthermore, no

breed effects were found in a previous study at the DRU

(Burridge et al., 1979).

Although fetal sex showed some statistical association

with precolostral antibodies when both groups were combined,

this possibly resulted from less fastidious postnatal

sampling of bull calves. Colostrum may have been consumed

by more bulls than heifers prior to sampling because less

attention is usually paid to postnatal care of bulls.

From a practical standpoint, knowledge of dam age or

parity probably would not assist in control or prediction

of in utero BLV infection. Economic justification for

efforts to control in utero infection seems unlikely with

a rate as low as 6.4%. Even though control may not be


possible, awareness of in utero infection is essential

in planning postnatal management of BLV transmission.


*H -.
O co

0 Or

C) -r-4

( o 0
*-d C) rC-
rd (U) r1

0 4


0 cl

O a
-d -H
6 rd S-

C3 >C4

*4 rd
,U -) t -
0 -H
r-o Cz (
0 () Q

04 Q) '-

4-4 0


> fd

CO *l C











'o 0 m in
T C N (N
m (N CN r-l

LO M m r-

(N- CN 0w

m or- No
Hl Hl

m0 0 -- CO
CO 0 co
Hl rl

/- m
o H

N 0 cy

En n 0-

M 0- (N

0 0 0
- -c- O.
5 O r>

(- 0 0C

PQ 4- 4
4 4J *0

o', 0-i 0
0 0 0 Q-I
ua Q- I -1

a) C *) 4
(0 0

P 414 4

0 0



a a

i mr

0 0

0 F- 0f
4-' 0 C:

0 0 0

O r
O .0

0 0 0

) > d
0o (

4-4 H) m

0 0 0

S0 C 0
1a rc$

Cd (D U)

4- +4- 4+

0) 0) 0)
o o4 04
a) a) a)
-c-i 4- .-
0u 0 0O
a) ) a
o ) .
5-C 54 5-
a) a) a

-Ic +. +


4 -

>i 4-
Cl I

4-0 0

Hd 0
0o -


5 -

I *H-



*H (
(0 .-I

4 0

0 2
O -H

0 0
O 4


O -H

0 0l

0 r.
0 3

r -4 -q

H 0




rF< 1-





tn S






Q) Q) C) C) > >1 >1 >
4J 4-) 4-i -3 a) ) C)

r-l rl r4l -4 l 4 4 4 4
0 0 0 0 (1) (U ) )

r, mi I. r-c ,,D () () m
rY 10 CN M^ Ln m CN r-

(N O ,-I (N o m ( N Lr,

.- -k
-KI -c ,N Co -K ---- P m
r O C Ci [- m0 0
Om o i o -i- c( N

La a a a tL) 1 -
b LA o ~-lo L 1
COOm lrbrrlr

4-I -H

-H 3


c c
H 03

T4 0

C 3

4 C

0 0

C)C0 i-
!O 0

CO 4
0 a0 r0l

3 tp a

>U C O) 4
*rl M0 r.

3r4 -HO)

z 0 C 4 (d

XI *H 03*
z0 0rl (U

4) r-l 0








J 4-

( -I

O >
r 4


0 *>
a fd




0 03




r )- i-




4-4 T
cl kaa)
a) 4-1

Hr U >1
H- 0 Q
0 ) 0) )

0 r a) 4 (U)
U0g H C on C U)

> 4







0 4

4) a(
u 1 Q
cl a
HcO 0

- r- D C

0o l- k0 -1

0 N 0 0

co In r-I o

O C4 0 0


0o -
o o 00 d

Mo o o

,.I ,- -

r- r 1 -1 r-


















04 -H
( >

4 E



O (
0 O
-r-4 -H

-P -4>

) ,-4
(0 0

0 0
en 4-)



0 )

0 0
4 0

e .3

3 >-i
) r-4

U 0

cc; r










4 -4


(0 0

z 0

M 0
4-H A
0 4J

a) >

04- i









a) a


o o

z 4-4

q 0

0 0


.H n

04 rd

4- 0
4 0Q

-K +






1. 0 *-
p ()

0 O C

3 0 4-r
jH > 4

r-4 C) O
> 00

0 > a) f

O a-

)0 C
4 1* 0

0 0 0

-4 04 04.
0- 0

04 ) 0 0
(a> *-4 >

F 4 4,-

0 z
at En 0
(0 > > 0
(1- 0 0
0 C)

4l u 0
0o 0

4 CCi

2 4- O4-

0U r-I (
P0 M 4-4
44J U 0 4J
4.) 0






> Q)

0 -r


z ?



t '




O r

C !







,r- -




r d

0 H

t .
0 o

4 H
o I+
0 -1

02 .
*c +



Before postnatal transmission of BLV can be studied,

criteria for serologic detection of BLV infection must be

established. This is straightforward for an animal which

seroconverts after a period of seronegativity. However,

during the period when colostral antibodies are detectable,

it is difficult to discriminate between infected and non-

infected calves strictly by the presence or absence of anti-

bodies. In order to more accurately define time or place of

infection, it would be helpful to detect infected calves

during the period that colostral antibodies are detectable.

This chapter will describe the decay of colostral BLV

antibodies in noninfected and infected calves and estimate

normal limits of decay in noninfected calves. It is shown

how these normal limits can be used to detect BLV-infected

calves during the first six months of life.

Materials and Methods

Data Analyzed

Titers were analyzed from calves which had colostral

antibodies on at least three observations, followed by no

serologic reaction for at least two months after the last

colostral antibody titer. These calves were assumed not to

be infected with BLV during the period when colostral anti-

bodies were detectable because, had the virus been present,

an active antibody response probably would have been ob-

served within two months after the last detectable colostral

titer (Mammerickx et al., 1980; Van Der Maaten and Miller,

1978b, 1978c). Furthermore, apparently to date, there is

no evidence that the presence of BLV appreciably alters the

rate of decay of colostral antibodies.

Titers also were analyzed from 14 calves which either

had precolostral antibodies that persisted throughout the

study or had antibodies persisting through eight months of

follow-up. These calves were assumed to have been infected

during at least part of the colostral-antibody period.

Only calves with titers which remained the same or declined

from previous levels were included in regression analysis.

All titers from infected calves were included in estimates

of ages at which BLV infection could first be serologically



Titers were analyzed using an estimated weighted re-

gression procedure (Swamy, 1971). All computations were

performed using SAS. A linear regression predicting the

logl0 of the inverse of BLV antibody titer at various ages

was determined by pooling intercepts and slopes from

individual calf least squares regressions. Each vector of

regression coefficients was weighted by the inverse of its

estimated covariance matrix. The mathematical model was

Y = a + bX + u

where Y (dependent variable) was logl0 of the inverse of

the end-point titer, a was the pooled Y intercept, b was

the pooled rate of decay of logl0 inverse titer, X (inde-

pendent variable) was age in days when the titer was ob-

served, and u was random error. Prediction bands (normal

limits) were calculated for the 90%, 95%, and 99% confidence

levels. A zero titer level was set at log10 (0.5) = -0.3.

Half-life of antibodies was estimated by

n log (10 ) a.
l1gl0 1

where a. and b. were the Y intercept and slope of the es-
1 1
timated regression line for the i calf and n = 130 calves.

A randomly selected set of 61 calves was used to ob-

tain preliminary estimates of a and b and normal limits.

The validity of these estimated normal limits of BLV anti-

body decay was tested using the remaining 69 calves as a

validation group. Since approximately 91%, 94%, and 98% of

the observations in the validation group fell within the

90%, 95%, and 99% normal limits established by the first

set of calves, the two sets were combined and a final pre-

diction model was derived based on all 130 calves.

Regression analysis was repeated for 14 infected

calves which had declining titers for at least three ob-

servations beginning at two days of age. Tests of hypothe-

ses of equal intercepts and slopes of colostral decay were

made using large approximate t statistics (Swamy, 1971,

p. 129).


For the 130 noninfected calves, the number of positive

antibody titers per calf ranged from three to eight. One

of these calves had an early titer as high as 1:256. Dura-

tion of colostral-BLV antibody titers ranged from 51 to 187

days with 50% of the calves serologically negative by 95

days of age (Fig. V-l).

The regression equation for all 130 calves was Y = 1.29

- 0.012X. The mean half-life of BLV antibodies was esti-

mated to be 27.1 days 1.2 days. The prediction line

crossed the zero titer level at 136 days of age, while the

90%, 95%, and 99% upper prediction bands crossed at 168,

178, and 200 days of age, respectively (Fig. V-2).

Serologic diagnosis of BLV-infected calves was made

between 2 and 180 days of age using the 90% limit and be-

tween 8 and 206 days of age using the 99% limit (Table V-l).

The regression line of BLV colostral antibody decay

for 14 infected calves was Y = 1.43 0.012X (Fig. V-3).

No differences were found between noninfected calves and

infected calves for slopes (p= 0.45) or intercepts (p=0.43).


Examination of individual calf regressions revealed

little deviation from linearity when titers were expressed

on the logl0 scale. The decision to set the level of a

negative titer at log10(0.5) was based on the two-fold

dilution scheme. Extrapolation below a titer of 1:1 can be

made with confidence because there is no reason to suspect

appreciable deviation from exponential decay below the

level detectable by AGID.

Some assumptions for estimated weighted regression

analysis (Swamy, 1971) could not be perfectly met in this

application. The distribution of log10 of inverse titers,

for example, cannot be assumed to be normal, particularly

as calves age, because titer values truncate at zero. Also,

the method of titer measurement results in a discrete de-

pendent variable which would not have an exact normal dis-

tribution. It is likely that with a sample of 130 calves

violations of these assumptions are not critical. Never-

theless, 69 calves were assigned randomly to a validation

set to test validity of the 90%, 95%, and 99% normal limits.

Since results showed little deviation from expectations, it

can be assumed that these limits were valid. Estimated

weighted least squares analysis was performed instead of

ordinary least squares regression because the assumption of

independence of deviations from overall regression does not

hold when repeated measures are made on each animal (Swamy,


It is well known that IgG is the major immune globulin

acquired from colostrum by the neonatal calf (Pierce and

Feinstein, 1965; Porter, 1972). These colostrum-derived

immunoglobulins originate from the maternal circulation

where they transfer to colostrum several weeks prior to

parturition (Brandon et al., 1971; Logan et al., 1973;

Smith, 1971; Sullivan et al., 1969). Since AGID using BLV

gp-51 antigen measures IgG1 exclusively (Matthaeus et al.,

1978), decay of colostral BLV antibodies is essentially the

decay of IgG The half-life of IgG1 in the bovine may vary

from 18 to 22.7 days (Brar et al., 1978; Dixon et al., 1952;

Logan et al., 1973; MacDougal and Mulligan, 1969; Smith,

1971). The longer half-life of 27.1 1.2 days found in

this study may have several explanations. Use of larger

and older animals in at least one study may account for a

shorter half-life, since younger and smaller animals may

metabolize IgG1 at a slower rate (Dixon et al., 1952).

Furthermore, the rate of decay of IgG1 may be dependent on

the incidence of diarrhea in calves studied (MacDougal and

Mulligan, 1969).

The 187-day duration of colostral antibody titers to

BLV found in this study was similar to that for infectious

bovine rhinotracheitis (Brar et al., 1978) and bovine viral

diarrhea (Brar et al., 1978; Kahrs et al., 1966; Kendrick

and Franti, 1974). In another study of 21 calves, colostral

antibodies to BLV were detected for only 154 days (Fischer

and Keyserlingk-Eberius (1980). Since AGID has not been

standardized, a discrepancy of this magnitude is not sur-

prising. A lower rate of diarrhea or consumption of a

higher quality or quantity of colostrum also could explain

the longer duration of antibodies found in our study.

The age at which infected calves could be detected was

variable within a given level of normal limits. This is not

unexpected because the critical level at which viral repli-

cation is initiated also is variable and dependent on both

neutralizing antibody and virus concentration (Driscoll

et al., 1977; Straub, 1978b). At the 90% prediction level,

infected calves were identified as early as 2 days and as

late as 180 days of age. At the 99% level, calves were

identified later, from 8 to 206 days of age, but fewer non-

infected calves were misidentified. The sensitivity of

this method varies inversely, whereas specificity varies

directly, with the level of precision (prediction).

Serologic detection of BLV-infected calves during the

period of colostral antibody decay could be used in eradi-

cation or control programs where it would be feasible to

serially dilute serum from calves less than six months of

age. Because typical levels of passive antibodies vary

from region to region and herd to herd (Naylor, 1979), a

decay curve might be defined for noninfected calves which

is characteristic of a herd or region. Since at least three

positive observations are required per animal to estimate

error variance, calves could be bled at weekly intervals

for the first month and at monthly intervals thereafter.

This technique of defining upper limits of colostral

antibodies also could be used in planning strategies for

herd immunization (Kahrs et al., 1966; Kendrick and Franti,

1974). Calves would be vaccinated at the earliest age that

passive antibodies would not interfere with immunization.

Other factors, such as herd disease prevalence, antigen dose

and type, and general health conditions, also should be

considered (Brar et al., 1978; Kahrs et al., 1966; Kendrick

and Franti, 1974; Uhr and Bauman, 1961).

Colostral antibodies decayed at the same rate from the

same level in infected calves as in noninfected calves.

This validated the assumption that presence of BLV did not

alter BLV colostral antibody decay. Use of calves with no

evidence of BLV infection for at least two months beyond

the last colostral titer, therefore, should provide valid

estimates of BLV colostral antibody decay.

Table V-l. Ages at which bovine leukemia virus (BLV) anti-
body titers from BLV-infected calves were above
normal limits of colostral titers of noninfected

Age (days) at which logl0 1/titer* was first
above the normal limits of
calf number 90% 95% 99%

J166t 2 2 178
J171 153 153 153
J183$ 151 151 151
Jl94Bt 169 169 169
J200Bt 38 73 206
H875t 180 180 180
H888 154 154 154
H913$ 155 155 155
H938 112 112 167
H949 2 2 8
H955t 160 160 160
H972 137 137 163
H973 67 67 102
H1002Bt 31 31 31
H1038Bt 27 27 144
H1079B 167 167 167

MEAN 106.56 108.75 143.00
SEM 16.54 16.07 13.22

Antibodies measured by agar-gel immunodiffusion using
glycoprotein antigen.

tCalves with precolostral antibodies to BLV.

tCalves with no detectable precolostral antibodies to BLV.

Calves which were not sampled precolostrally.


,) 80.

Z 60-


> 20.

- - -


20 40 60 10 1 l 10 1? 2)t


Fig. V-1. Percentage of 130 calves with colostral anti-
bodies to bovine leukemia virus (BLV) as a
function of age.

1. _
> L
'- 15

Z 0.9

> 0 0.6




Fig. V-2.

0 0


----99% LIMITS
- 95% LIMITS
...... 90% LIMITS




0AooAoA '0 00


Prediction line of the decay of colostral anti-
bodies to bovine leukemia virus (BLV) in 130
calves with no evidence of infection with BLV.
Upper normal limits of the prediction line are
indicated for the 90%, 95%, and 99% confidence
levels; o=values of logl0 1/titer falling
above the 90% normal limit from calves not in-
fected with BLV; A=values of logl0 1/titer from
calves infected with BLV in utero or sometime
prior to six months of age.





1.5 -..J4




0 a 0

* ^s* *

*** ** *** *

, 00 ~

* 0

0 0

20 40 6 0 100 120 1
20 40 60 80 100 10 140

160 1 0 200 210


Fig. V-3.

Prediction line of the decay of colostral anti-
bodies to bovine leukemia virus (BLV) in 14
calves infected with BLV in utero or sometime
prior to six months of age; =values of log10




co _











Several factors have been examined individually for

association with prevalence rates of BLV infection in previ-

ous studies. Rates have been reported to be unrelated to

BLV-status of the dam (Hofirek, 1980; Olson et al., 1978;

Valikhov, 1978), consumption of colostrum from BLV-infected

cows (Ferrer et al., 1976; Ferrer and Piper, 1978; Piper

et al., 1975, 1979), or breed (Burridge et al., 1979).

Other studies have suggested that an association exists

between prevalence rates of BLV infection and breed

(Burridge et al., 1981), BLV status of the dam (Baumgartener

et al., 1978), type of colostrum consumed (Seger and Morgan,

1977), and age (Burridge et al., 1979, 1981; Chander et al.,

1978; Evermann et al., 1980; Ferrer et al., 1976; Hofirek,

1980; Huber et al., 1981;Mammerickx et al., 1978a, 1978b;

Marin et al., 1978; Olson et al., 1973; Parfanovich et al.,

1978; Piper et al., 1979). Because age is associated with

BLV prevalence rates, it is important to examine for the

age-specific effects of factors on rates of infection.

The objective of this chapter is to describe age-

specific rates of detection of BLV infection and to test

for differences between rates for cattle identified by

various maternal and management features.

Materials and Methods

Cattle Studied and Criteria for Detection
of BLV Infection

Cattle studied were those described in CHAPTER III.

Calves known to be infected in utero (CHAPTER IV) were ex-

cluded from examination of postnatal infection. Detection

of BLV infection was made when BLV-gp antibodies first ap-

peared following a period of seronegativity. In calves

showing evidence of colostral BLV antibodies, e.g., a declin-

ing titer during the first six months of life, the time at

detection of BLV infection was determined from the model of

colostral antibody decay (CHAPTER V). A calf was detected

as infected when titers first began to increase or when a

titer fell outside the 95% normal limits of colostral anti-

body decay, whichever occurred first. Because the precise

moment of infection could not be measured, the date of de-

tection was used to approximate this time, assuming a uni-

form lag period from infection to detection in all cattle.

A BLV-infected dam was defined as one which had anti-

bodies to BLV during gestation or within one month follow-

ing parturition. A calf was judged to have consumed colostrum

from a BLV-infected cow if the calf had detectable anti-

bodies to BLV in the first postcolostral sample and was

precolostrally negative or if it had postcolostral antibody

titers that declined with age.


Data were analyzed using survival regression tech-

niques (Cox, 1972; Taulbee, 1979) available in the PHGLM

program of SAS. Life-table calculations and graphics were

performed using the statistical program P1L of BMDP. Fail-

ure time was defined as the age in days when an animal was

detected as infected with BLV. Censored time was the age

when an animal died, left the herd, calved, or when follow-

up was terminated on September 30, 1981, whichever occurred


The number of cattle at risk of being detected as in-

fected in an age interval ti, defined for graphical purposes

as 30 days, was

r. = n.
1 1 2

where n. was the number of animals entering the interval and

c. the number censored in the interval. Conditional prob-

ability of detection in an interval (given that detection

did not occur in the previous interval) was

q 1

where d. was the number detected as infected in the interval.
It was assumed that risk of detection remained constant

during an interval. The conditional probability of an

animal not being detected (surviving) in the ith interval


pi = 1 q

The cumulative proportion remaining undetected (also known

as cumulative survival or survival rate) to the beginning
of the i interval was

Pi = Pi-Pi-

where P =1. Hazard rates (also called failure rates or

force of morbidity) were estimated at interval midpoints

and were defined as the quantity

Xi = 2qi/hi(l+pi)

where h. was the width of the i interval. This rate was
a linear interpolation estimate based on a life table rep-

resentation of a survival curve.

Standard errors of cumulative survival functions and

hazard rates were calculated as described in the P1L pro-

gram of BMDP for all animals born between July 1979 and

June 30, 1981.

Relationships between hazard rates and other variables

were assessed using the model of the hazard at time t, as

previously described (Cox, 1972) because it adjusts for

censoring in the data. This model is

A(t,zi) = exp(z.i)A(t) ,

where 8 is a vector of unknown parameters and z is the vari-

able under study. The variable z can take on indicator

values (e.g., 1 =positive dam and 0 =negative dam) or

continuous values as in dam age. Implicit in this model

is the assumption that hazard rates for individuals with

different values of the covariate are in constant ratio

over time, regardless of the underlying hazard rate

(Taulbee, 1979). An obvious violation of the proportional

hazards assumption occurs when survival curves cross. This

often is the case with actual data, and it is possible to

examine changes in hazard rates with age using an alterna-

tive function described by Taulbee in which a new variable,

zt, is defined. This allows the ratio of hazard functions

to differ and permits examination for age-by-factor inter-

actions. Conclusions based on this test, however, must be

extremely tentative, since the differences detected as

significant are unknown.

Correlation coefficients were calculated using the

CORR program of SAS to describe relationships between the

various factors investigated. To examine effects of factor

interactions on survival, correlation coefficients were

computed from the variance-covariance matrix of parameter

estimates when all factors were included in the survival

model. High negative correlations between regression pa-

rameters for two factors were interpreted as a reduction in

the effectiveness of one factor in explaining survival when

the role of the other factor on survival became more

important. High positive correlations identified factors

which were jointly effective in explaining survival.

Factors examined were dam age (months), dam parity

(0 =primiparous, 1 multiparouss), dam BLV status when the

animal under study was born (1 =positive, 0=negative),

type of colostrum consumed (1 = from a BLV-infected cow,

0=not from a BLV-infected cow), and breed (1= Hostein,

0 =Jersey).


A total of 473 live calves entered the survival analy-

sis at birth. Of these, nine were detected as infected in

utero (CHAPTER IV) and 54 infected postnatally. Two ani-

mals were followed through 780 days of age (Table VI-1).

Estimates of prevalence rates of infection increased from

2.3% at birth to 63% at 27 months of age (Fig. VI-1).

Survival rates (Fig. VI-1), hazard rates (Fig. VI-2), and

proportion of animals detected as infected for various age

groups (Fig. VI-3) demonstrated four possible age-related

stages of detection. The first stage, prenatal infection,

has been characterized previously (CHAPTER IV). A second

stage appeared as increased hazard rates during the first

six months of life. This was followed by a 10-month period

of sporadic detection. The fourth stage, beginning at 16

months of age, was characterized by sharply increasing

hazard rates of detection through 27 months of age.

Results of statistical analyses using Cox's procedure

are presented in Table VI-2. No differences were found in

hazard rates when variables were examined separately without

the age covariate. However, significant age interactions

were suggested for all factors studied after inclusion of zt

and use of the test by Taulbee (Table VI-2). Graphically,

factor-by-age interactions were indicated for calves con-

suming BLV-positive colostrum (Fig. VI-4), for dam status

(Fig. VI-5), for breed (Fig. VI-6) and for dam parity

(Fig. VI-7). Lower rates of detection of BLV infection

were observed in young calves from infected dams and in

calves consuming colostrum from infected cows. Higher rates

of detection were observed in older Jerseys than in Holsteins

of the same age. Calves born to primiparous dams appeared

to have lower rates of detection through 18 months of age,

after which they began to experience higher rates of detec-

tion until 27 months of age when rates were similar.

Significantly high factor correlations were found be-

tween type of colostrum consumed and dam status (p=0.0001)

and between dam age and parity (p =0.0001), and low, but

significant, correlations were found between dam parity and

dam status (p =0.0001), between colostrum and parity (p=

0.001), and between colostrum and dam age (p =0.001) (Table

VI-3). High, negative correlation coefficients of estimated

parameters of survival were found only for dam age and

parity (r=-0.97) and for colostrum and dam status (r=

-0.86) (Table VI-4).


Throughout this study increasing or persisting antibody

titers to BLV, as measured by AGID-gp, were interpreted as

evidence for BLV infection. Because this method indirectly

tests for presence of virus by an immune response, some as-

sumptions regarding sensitivity and specificity of the

diagnostic criteria should be made.

The BLV-gp antigen appears to be unique, since no cross

reactivity with proteins of other oncoviruses or retro-

viruses has been demonstrated (Ferrer, 1972; Ferrer et al.,

1975; Kaaden et al., 1977; McDonald and Ferrer, 1976;

McDonald et al., 1976). A positive serologic response to

BLV-gp, therefore, can be considered specific for BLV.

Comparisons of several serologic tests have demonstrated

that AGID-gp is a relatively sensitive serologic test (Burny

et al., 1980). These results are corroborated by experi-

mental studies in which a serologic response following BLV

inoculation was consistent (Miller et al., 1972; Schmidt

et al., 1976; Van Der Maaten and Miller, 1978b; Van Der

Maaten et al., 1981a) and persistent (Miller and Van Der

Maaten, 1976b). Similarly, BLV has been readily detected

from seropositive but not seronegative animals following

natural infection (Ferrer et al., 1976, 1977a, 1977b;

Olson et al., 1973; Piper et al., 1979).

There is no evidence for recovery from BLV infection.

On the contrary, seropositive cattle older than six months

appear to maintain detectable antibodies throughout life

(Chander et al., 1978; Kaaden et al., 1978; Tabel et al.,

1976). In a study to examine vaccine efficacy, inactivated

BLV produced only a transient serologic response, and virus

was not isolated (Miller and Van Der Maaten, 1978b). In

that study two doses of killed virus were required to induce

a serologic response. This suggests that persistent anti-

body titers only result from active, persistent infection

and that any transient serologic response resulting from

exposure to a killed virus is rare.

Until recently, it could not be stated that absence of

either virus or antibody implied absence of specific BLV

sequences in host cell genomes. Studies now suggest that

such sequences are not to be found in cattle without BLV or

specific BLV antibodies (Callahan et al., 1976; Deschamps

et al., 1981; Kettmann et al., 1976).

A final demonstration of the sensitive nature of AGID

using gp-51 lies in results of BLV-eradication programs

based upon mass screening with this test. Results of these

campaigns indicated that the virus was efficiently removed

from aherd by test-and-slaughter methods (Mammerickx et al.,

1978a; Schmidt et al., 1978, Straub,1978b). If AGID were

not an efficient discriminator of infected and noninfected

cattle, it is unlikely that eradication efforts would have

met with such rapid success. Use of AGID-gp to detect BLV

infection, therefore, is justified, particularly in a design

in which animals can be followed for several months and

persisting titers can be observed.

Survival analysis and life-table methods are well es-

tablished techniques in epidemiology and actuarial science

(Gross and Clark, 1975; Lilienfield, 1976). Recently these

methods have seen increased application in veterinary

epidemiology (Cobo-Abreu et al., 1979; Hird et al., 1975;

Schwabe et al., 1977). Survival analyses are especially

appealing in prospective studies where animals lost to

follow-up can be considered in computations up to the time

of censoring, assuming failure rates similar in censored

and uncensored animals.

Data presented in this study are particularly suitable

for examination by survival methods for two reasons. First-

ly, recovery from BLV infection is unknown and animals can

be defined as truly failed once infection is detected.

This is analogous to death in traditional applications.

Secondly, since these methods incorporate the dimension of

time, either by comparing distributions over time or by

using time as a dependent variable, they are appropriate

in studies of BLV infection because prevalence rates are

known to be associated with age (Burridge et al., 1979,

1981; Chander et al., 1978; Evermann et al., 1980; Ferrer

et al., 1976; Mammerickx et al., 1978a, 1978b; Marin et

al., 1978; Olson et al., 1973; Piper et al., 1979).

Cox's model was used to test estimated parameters of

a survival function where the underlying assumption was

that hazard rates were proportional (Cox, 1972; Taulbee,

1979). The model tends to force nonparallel curves to a

parallel model. Thus, interaction, characterized by cross-

ing or bellying curves, could not be examined. Inclusion

of the zt variable removes dependence on proportionality and

allows for tests of interactions with age (Taulbee, 1979).

Estimated age-specific prevalence rates were calculated

as one minus the proportion remaining negative for BLV in-

fection. The prevalence rate estimate for 18-month-old

cattle of 25% was lower than that of 45% described from a

cubic polynomial curve fitted to data collected from the

same herd for the years of 1975 through 1977 (Burridge

et al., 1979). Transmission rates of BLV infection, there-

fore, appeared to have declined over the past four to five

years, at least in animals up to 18 months of age. This

may suggest variations of prevalence rates in yearly birth

cohorts, as found in other prospective studies (Huber et al.,

1981; Wilesmith et al., 1980). In those studies, it was

observed that cattle born in a given year experienced a

different life-time prevalence rate than did animals born

in some other year. Furthermore, the prevalence rate

within a cohort did not change appreciably beyond two

years of age. A similar situation may have occurred in

this herd due to management of Holstein heifers. Prior to

1977, bred Holstein heifers were kept in the dry herd until

shortly before calving. In the past few years, these

heifers have been managed separately from the dry herd,

whenever pasture space permitted, in order to provide them

with a better plane of nutrition. In this study Holstein

heifers were exposed to the dry herd for a relatively short

time, January through April 1981, which could have limited

BLV transmission and resulted in the lower prevalence rates.

Inspection of the overall life table revealed heavy

censoring during the first two to three months. This was

due to sales of bull calves and to neonatal mortality.

Early postnatal detection of BLV infection, constituting

the second detection phase, could have been represented by

calves infected in utero and not sampled precolostrally or

infected too late in gestation for antibodies to appear by

birth, as previously suggested (Van Der Maaten et al.,

1981b). Evidence that postnatal infection occurred during

this period was demonstrated by detection of infection in

calves born to noninfected dams.

The length of this period of detection would not cor-

respond to the length of an underlying infection period

because of the seroconversion lag time (CHAPTER VIII) and

possible repression of viral expression by colostral anti-

bodies. Inhibition of release of BLV by antibodies has

been demonstrated in vitro (Driscoll et al., 1977) and also

suggested in studies of colostral antibodies (Van Der Maaten

et al., 1981a). Antigenic modulation or inhibition of P

proteins (Sissons and Oldstone, 1980) may explain such a

phenomenon. The process of modulation is believed to

involve a binding of viral glycoprotein on the cell membrane

by circulating antibodies. Glycoprotein surface markers

become masked from recognition by K cells or other cells

invoking an antibody dependent cytotoxic response. A

failure to recognize virus-bearing lymphocytes by the

immune system renders them resistant to immune damage. If

antibodies are removed, for instance through natural at-

trition of gammaglobulins, the lymphocyte would regain the

ability to shed virus. Inhibition of P proteins may also

be involved in repressing virus expression by inhibiting


A further delay in detection, beyond the two- to three-

month seroconversion period and one-month sampling interval

could have resulted from early sampling-design flaws. As

much as a two-month delay could have occurred as a result

of the three-month sampling interval used at the beginning

of the study. It is likely, therefore, that infection

represented by this phase of detection occurred in the

first two to three months of life or even prenatally.

Several factors may have contributed to postnatal

transmission in the first two to three months. Calves may

have become infected at parturition when a calf born to an

infected dam ingested maternal blood from uterine or vaginal

tears (Van Der Maaten et al., 1981b). Rates of detection

in calves from infected dams, at least during the first

three months, however, were not different from rates in

calves from noninfected dams. Such a similarity of rates

may be an artifact due to delayed expression of the virus

by colostral antibodies. Consumption of colostrum with

BLV antibodies, therefore, may complicate control efforts

by masking early infection.

Type of calf management also may have influenced BLV

transmission during the first few months of life. Most

calves were placed inside the calf barn where they could

have been exposed to infected calves in adjacent pens.

Transmission may have resulted from ingestion or inhalation

(Van Der Maaten and Miller, 1978c) of saliva (Ressang et al.,

1980), urine (Gupta and Ferrer, 1980), blood (Mammerickx

et al., 1980; Van Der Maaten et al., 1981a), or lymphocytes

(Miller and Van Der Maaten, 1978b; Van Der Maaten and

Miller, 1978b, 1978c, 1981a). Animals were also tattooed

while in the barn, introducing the opportunity for iatro-

genic transmission. Other calves occupied reasonably iso-

lated pens outdoors, and they may not have experienced the

same risks of infection as indoor calves. Risks for animals

in various locations are examined further in CHAPTER VIII.

A third relatively quiescent phase of detection of BLV

infection between 6 and 15 months of age coincided with

movement of animals from densely occupied paddocks to

fields and pastures (APPENDIX D, Fig. 7). It also was noted

that vaccinations for other infectious diseases were given

either shortly before or during this period. Iatrogenic

transmission of BLV, therefore, does not appear to have

been of major importance in transmission because rates of

detection actually declined during this period. A more

detailed examination of detection rates and vaccination pro-

cedures is presented in CHAPTER VII.

A sharp increase in detection rates typified the fourth

and last observed phase which began after 15 months of age.

The only management interventions preceding the increased

rates were insemination and movement of bred heifers to the

dry herd. Infection as a result of close physical contact

has been suggested as the major mode of BLV transmission

(Dechambre et al., 1968; Maas-Inderwiesen et al., 1978;

Straub, 1971, 1978a; Miller and Van Der Maaten, 1978a;

Wilesmith et al., 1980). Other reports mentioned similar

observations or seroconversions following exposure to in-

fected animals (Ferrer et al., 1976; Piper et al., 1979).

However, in our study not all bred heifers entered the dry

herd, permitting a comparison of risks of two types of

heifer management in CHAPTER VIII.

It is not likely that transmission resulted from in-

semination, since the exogenous nature of BLV excludes

transmission via gametes. This has been confirmed by

observational studies which did not find increased infec-

tion rates in animals sired by BLV-infected bulls (Baum-

gartener et al., 1978) and by experimental studies which

failed to transmit BLV with semen from BLV-infected, but

otherwise healthy, bulls (Miller and Van Der Maaten, 1979;

Ressang et al., 1980). Furthermore, the time from breeding

(13-16 months of age) to peak rates of detection (18-21

months of age) was longer than suggested by experimental or

natural transmission studies (Gentile and Rutili, 1978;

Mammerickx et al., 1980; Straub, 1978b; Van Der Maaten and

Miller, 1978a, 1978b). Unfortunately, an effect of in-

semination on transmission could not be examined further

because BLV status of sires was unknown and analysis would

have been confounded by the presence of only bred heifers

in the dry herd.

In the examination for associations between hazard

rates and dam age, dam parity, dam status, breed, and colos-

trum type using Cox's model, no significant main effects

were observed. Results of Taulbee's alternative procedure,

however, suggested significant departures from proportional

hazards. The sizes of the differences detected by this al-

ternative method are not known, and they may be too small

to be of practical importance in the control of BLV in-

fection. Conclusions regarding the practical significance

of interactions, therefore, should be very tentative, but

discussion can be based on inspection of survival curves.

Graphical examination is meaningful because crossing of

survival curves cannot occur without crossing of hazard

curves, thus allowing for visual inspection of factor-by-

age interaction.

Interactions between age and dam status and between

age and colostrum have already been discussed and indicated

that calves which consume colostrum from an infected cow or

which were born to a seropositive dam were protected from

detection but not necessarily infection, at an early age.

These two factors would be expected to have similar curves

because of the high negative correlation of their estimated

survival parameters. Although several factors were sig-

nificantly correlated with each other, only dam parity and

dam age, and dam status and colostrum were correlated when

hazard rates of detection were considered.

An interaction of breed by age was suggested when old

Jersey heifers were observed to have higher rates of detec-

tion than old Holstein heifers, but at young ages curves were

similar. It has been reported that Jerseys have low levels

of circulating gammaglobulins (Logan et al., 1981), but

this would not necessarily explain increased susceptibility

only in old Jersey heifers. A more reasonable explanation

for lower survival rates in old Jersey heifers is that they

remained in the dry herd until shortly before calving.

Holstein heifers, on the other hand, were removed from the

dry herd on about May 1, 1981, and Holsteins bred 45 days

prior to that time were never in the dry herd. The breed-

by-age interaction, therefore, was due most likely to

longer exposure of Jerseys to older, infected cattle in

the dry herd.

There has been no mention made in the literature of

dam parity or dam age effects on subsequent BLV infection

of progeny. Crossing of survival curves of animals from

primiparous and multiparous dams observed in this study

suggests an age-by-parity or age-by-dam age interaction.

Whether the crossing of these curves represents expected


variation, especially with small numbers at older ages, or

a real effect cannot be stated. No other explanation can

be provided.

com -



















: -I

4- 4






0 Oq










L(N min m

0 .o o0

LOn -I Ln r-
000 CO0


" i-i Ln %10 .-i 1, 'o 0 0 -i 0 (N






Nr Co r-I
r-I (N

(N r-4

Lfn 00 r-4 0 O)

( r- r- N

CN ,-I r- r-
1-4 -4 1-4 r-i r-4

N Ln co

-I0 O0 N
O 000

co Lfn 1- (N ^r rN
0 on -I cm (N


Om, CO 00
-N -4 -4



















*i N 0 m4

H m


m m 1 M r-
H4 HH 04 H
00000; ( C

v i LnA nC4 0
n N 1-1 r- m 0
m o NT 0 ; v M Ln
o r-o0 -o0 0 o
O r~~
mO f Or

000000 0000 00000000

m 0r 00 oo


'. C4 0D C4' N L 0C4 'D
0N co LAn c LoA OI 4rn oo
Ln 04 tL 04 LA k.0 N- '. 0
O mo Loin z1 m


0 co
1-- o

O 04 H O A m L cN 0C 0 -H 0

C H-- 0N
-lH l- r-

H- 0N M M LAn LA 00 CM

Nr Ln m 0) CN 0 m m H (v)
0' 2 co '. w0 LA O


01 0 00 0 CO






" "n


U U) LA C> \ 0 n
r r- r- 113 in L o
"1-4 0 .
> -I o 0 0 0 0

0 o

S 0 o 1 1 LA o

0 H
i4 .- 0 L n 0

O O o O O O4
0 n O o o o o
& > O O O o O
0 0 0 0 0 0

0 (n
5 -4

s C) C 0
o k o

--4 1.1
U) C 0 0 0
4-) >1 0 0 0 0 0

a 0 0 0 0 0

> 0

I -A

> zq

*-H 0 '- I r-I r-l ,-1

S0 0 CN 0 0 0
U)4 44U
i0 0

0I 0C C

4- O 2
- (U0 *) U)
S(q < ( C O )
5- E -4 U )
M (i (d- 0 5- (
E- > Q m Q U m

9 II








c c











r-- C





.i 0

*o A

0 !4

5 ~A

4.1 4-
-H H

1 --



z -,

C) z

rd rZ
o 4
O r

a) o
o --

en enr
C, (U



o r

o C

Table VI-4. Correlation coefficients of estimated parame-
ters of hazard rate functions for the detection
of bovine leukemia virus infection.


Variable Dam status Dam age Colostrum Breed

Dam parity -0.002 -0.973 0.027 0.087

Dam status -0.027 -0.857 -0.021

Dam age -0.036 -0.052

Colostrum -0.037



CA 80-

u J
Z 60- /

>I /


I I I I I I 1 1 1 1 I I I 1 1 1 I 1 I i
3 6 9 12 15 18 21 24 27

Fig. VI-1. Estimated age-specific cumulative survival
(%) and prevalence rate (%) of detection of
bovine leukemia virus infection.




U .7-

-" .6-
-z .5

oi u.
0 0
< .3-


3 6 9 12 15 18 21 24

Fig. VI-2. Age-specific hazard rates of detection of
bovine leukemia virus (BLV) infection.



Z 10


9 12 15 18

Fig. VI-3.

Age at detection of bovine leukemia virus (BLV)
infection for 63 BLV-infected animals.

.9 ------------CALVES CONSUMING

Z .8

^ .7
Z c

2 .6

Fig. VI-4. Cumulative proportion of calves consuming colos--

antibodies remaining undetected for BLV infec-
tio .n.


100 200 300 400 500 600 700 800


Fig. VI-4. Cumulative proportion of calves consuming colos-
trum with bovine leukemia virus (BLV) antibodies
and of calves not consuming colostrum with BLV
antibodies remaining undetected for BLV infec-




Fig. VI-5.

Cumulative proportion of animals from bovine
leukemia virus (BLV)-infected dams and from
noninfected dams remaining undetected for
BLV infection.


Fig. VI-6.

Cumulative proportion of Holsteins and Jerseys
remaining undetected for bovine leukemia virus
(BLV) infection.

9- ----BORN TO
0 .8-

Z .7-


I-1 I I I I I I I I I I I-
-02 \

100 200 300 400 500 600 700 800

Fig. VI-7. Cumulative proportion of animals born to primi-
parous cows and of animals born to multiparous
cows remaining undetected for bovine leukemia
virus (BLV) infection.



A fundamental consideration in the epidemiology of in-

fectious diseases is the examination for temporal patterns

of occurrence of a disease. Seasonal or secular trends in

incidence rates may suggest modes of transmission, causal

associations, and/or etiologic agents. If, for example, a

hypothesized factor in disease transmission demonstrates

seasonal variability, and a statistically significant season-

al cycle of disease exists in phase with the hypothesized

factor, then supporting evidence favoring involvement of

the factor in transmission has been demonstrated. On the

other hand, if no seasonal occurrence of disease can be

demonstrated, or it is not in phase with that of the factor,

then the hypothesis of factor association should be rejected.

Interest in possible vector-borne transmission of BLV

has arisen from evidence for arthropod involvement in the

spread of another retrovirus, equine infectious anemia virus

(Fischer etal., 1973; Hawkins et al., 1973a, 1973b). Re-

ports based on experimental and observational data have

suggested the involvement of arthropod vectors in BLV trans-

mission (Bech-Nielsen et al., 1978; Ohshima et al., 1981).

Another study, however, revealed higher rates of infection

during winter months, coinciding with crowded, indoor hous-

ing conditions rather than exposure to large numbers of

flies (Wilesmith et al., 1980).

In this chapter, the hypotheses tested are as follows:

(1) age-specific rates of detection of BLV infection do

not differ with respect to month of birth; (2) rates of

detection of BLV infection show no significant seasonal

pattern associated with the seasonal frequency of potential

arthropod vectors; and (3) no other seasonal pattern of

rates of detection exists.

Materials and Methods

Month of Birth

Cox's hazard model described in CHAPTER VI was used to

examine for an association between month of birth and sub-

sequent BLV infection. Month of birth was denoted by z in

the model. Detection of BLV infection was based on cri-

teria presented in CHAPTER VI.

Month of Detection of BLV Infection

Monthly incidence rates of detection of BLV infection

were computed to estimate seasonal patterns of BLV infec-

tion rates. Monthly incidence rates of detection were cal-

culated as the number of animals detected during the month

divided by the number of animals at risk of being detected

in that month. The criteria for an animal to be at risk for

any month of follow-up were (1) the animal must not have

been detected previously; (2) the animal must have been

present for at least half of a month; (3) an animal born in

the last half of a month was considered only in subsequent

months; and (4) an animal remaining undetected for more

than one year would be considered also for each repeated

month. For example, if an animal were present during

January 1980, and January 1981, it would be counted as

two animals-at-risk during the month of January. Those

animals detected as infected during a month also were

counted in the at-risk denominator.

Possible differences between the 12 monthly incidence

rates were examined using a Chi Square test with 11 degrees

of freedom. In order to determine if monthly incidence

,rates followed a simple harmonic trend, a test for goodness-

of-fit of the data to a simple harmonic curve was performed

using a Chi Square statistic, as previously described

(Walter and Elwood, 1975). The test for a simple harmonic

trend of incidence rates, however, could not be pursued
because a high Chi Square value (X = 17.5, 0.05 (11)
indicated that these rates were inappropriately described

by a simple harmonic curve.

An alternate, nonparametric approach to the seasonality

of events was used to examine for six-month trends in in-

cidence rates (Hewitt et al., 1971) Rates were computed

only for animals over 12 months of age. The tested

hypothesis stated that six-month cumulative ranks of inci-

dence rates from May through October, the fly season, were

not greater than those for the 11 other six-month periods.

A second hypothesis was tested for no differences in cumu-

lative ranks of incidence rates for any nonspecified six-

month period.


Month of Birth

No significant differences were found between hazard

rates of detection of BLV infection for animals born in

each of the 12 months (p= 0.24). The proportion of animals

infected was highest in animals born in August, September,

and October (Fig. VII-1).

Month of Detection of BLV Infection

No significant differences were found between the 12
monthly incidence rates (X(II) = 9.81, 0.4
The numbers of animals detected as BLV infected and at

risk of detection for each of the 27 months of observation

are presented in Table VII-1. Monthly incidence rates of

detection for animals of all ages ranged from 1.1% in

January and April to 3.0% in June (Fig. VII-2). For ani-

mals over 12 months of age, cumulative ranks of monthly

incidence rates between May and October were not signifi-

cantly greater than those for any other six-month period

(p= 0.20) (Table VII-2). Cumulative ranks of monthly in-

cidence rates were highest for the six-month period be-

ginning with March and were lowest for the period beginning

with September. The probability of cumulative ranks was

lowest for the period beginning with March, but this was

not significant without an a priori commitment (p= 0.13).


Although no statistical significance could be placed

on differences between birth-cohort hazard rates, calves

born between August and October constituted 65% of the in-

fected animals, even though only 42% of calves were born

during that period (APPENDIX C). This was not surprising

for several reasons. The density was highest during the

period in which these calves were in the calf barn

(CHAPTER III). In a crowded situation, transmission may

occur either by aerosol (Van Der Maaten and Miller, 1978c)

or by contact (Dechambre et al., 1968; Maas-Inderwiesen

et al., 1978; Miller and Van Der Maaten, 1978a; Straub,

1971, 1978a; Wilesmith et al., 1980). Also, since these

animals were the oldest in the study, they had a greater

chance of entering the dry herd where there may have been

an increased risk of infection due to contact with older,

BLV-infected cattle. Calves born during the winter and

spring were too young to be bred and enter the dry herd by

May 1981. The proportion of calves born in the spring and

subsequently becoming infected would be expected to be low

because few cows calved at that time (APPENDIX C).

Two assumptions were made in the analysis of monthly

incidence rates of detection of BLV infection. One was

that animals were considered independent within and among

months followed. This is reasonable because exposure to

a possible BLV-carrying insect and susceptibility to BLV

infection were not likely to be related to any previous


The other assumption in examination of monthly inci-

dence rates was that any pattern of detection would be

synchronized with that of infection and would consistently

follow infection by two to three months. This assumption

may not be valid for calves less than six months of age.

Infection may occur in utero or shortly before birth but

before acquisition of passive protection from colostral

antibodies. In such cases, detection of infection may not

be made for six months or more because virus repression may

be dependent on both antibody and virus concentration

(Driscoll et al., 1977; Straub, 1978a). It is believed,

however, that viral expression would appear within 12

months of infection (Van Der Maaten et al., 1981a).

Analysis of possible vector transmission also may have

been confounded by calfhood management. Since the calving

season occurred during the summer and early fall, animal

density in the calf barn was highest during late summer and

fall. Seroconversion of animals several months later could

be interpreted as infection resulting from close physical

contact rather than from arthropod vectors. Furthermore,

the risk of infection for calves in the calf barn may not

have been as high as for those over one year of age. Flies

found in the calf barn were not usually the biting type,

and sporadic insect control was practiced, which limited

exposure to potential vectors. Colostral antibodies also

may have provided some resistance to infection from small

doses of virus introduced by an insect (Van Der Maaten

et al., 1981a).

The situation with cattle more than 12 months of age

was quite different. Biting flies were present on these

animals throughout summer months, and no fly control was

practiced. Prevalence rates of infection also were higher

in these animals (CHAPTER VI), increasing the chance that

a fly would first feed on an infected animal. Cattle over

one year of age, therefore, were analyzed separately as they

constituted a more susceptible and exposed population to

possible vector-borne transmission and because infection

was more easily detected and analysis was not confounded

by management factors.

Results of the Chi Square test indicated that there

were no significant differences between monthly incidence

rates. Important seasonal trends, however, may be present

without apparent rate differences between months (Walter

and Elwood, 1975). The method proposed for analysis of

seasonal trends in incidence rates (Edwards, 1961; Walter

and Elwood, 1975) was particularly appealing because season-

al frequencies of potential arthropod vectors in temperate

and subtropical climates generally resemble simple harmonic

curves which peak during summer months (Beck, 1958; Blanton

and Wirth, 1979; Bohart and Washino, 1978; Jones and Anthony,

1964; Khalaf, 1969). Their analysis could not be pursued

because the incidence rates were inappropriately described

by a simple harmonic curve. The goodness-of-fit test in-

dicated that incidence rates did not resemble a simple har-

monic curve and no further tests for significance or maximum

amplitude of a curve should be performed. This analytic

method also could not be used to test rates in older animals

because the recommended sample size is not less than 50

(Walter and Elwood, 1975). Instead, a nonparametric method

appropriate for small samples was used to examine for ex-

cessive rates over six-month periods (Hewitt et al., 1971).

Difficulties with Hewitt's alternative procedure have

been its lack of power and inability to estimate parameters

of an harmonic curve (Walter and Elwood, 1975). In spite

of these restraints and the few infections detected, a

marked tendency for maximum risk of detection of infection

was revealed between March and August (p= 0.13), the six-

month period beginning with March. Had an a priori hypothe-

sis been specified for that particular yearly segment, there

would be reason to believe that an excessive rate of de-

tection had occurred during that period (p= 0.01). This

interval would coincide with infection resulting from

exposure to animals in the dry herd between January and May

1981. If vectors had played a major role in transmission

and assuming peak infection around July, then maximum

cumulative ranks of incidence rates of detection should

have been observed between May and October. This was not

the case, as indicated by acceptance of the hypothesis of

no increase in incidence rates for that six-month period.

Results of this study are consistent with those from

a study in which increased rates of seroconversion were ob-

served after cattle grazed together on summer pastures

(Onuma et al., 1980). In that study, arthropod-borne

transmission could not be examined because analysis would

be confounded by congregation of cattle only during the

vector season. In another study, two groups were compared

for rates of infection during winter and summer months

(Bech-Nielsen et al., 1978). Rates observed for summer

and winter groups, 4/7 and 1/7, respectively, could have

occurred by chance (Fisher's exact p=0.13). A study of

incidence rates of BLV seroconversion in a large dairy

found higher rates in winter and spring, implying trans-

mission associated with crowded housing (Wilesmith et al.,

1980). None of these studies examined monthly rates for a

complete seasonal cycle. Future studies of seasonal patterns

of disease should be based on at least two years of monthly

cohort data in an attempt to remove as much year-effect as