Reprod Dom Anim 39, 429–433 (2004)

Ó 2004 Blackwell Verlag, Berlin

ISSN 0936-6768

Timing of Early Foetal Loss for Single and Twin Pregnancies in Dairy Cattle
F López-Gatius1, P Santolaria2, JL Yániz2, JM Garbayo3 and RHF Hunter1
1
Department of Animal Production, University of Lleida, Lleida, Spain; 2Department of Animal Production, University of Zaragoza, Zaragoza,
Spain; 3Department of Animal Production, Diputación General de Aragón, Zaragoza, Spain

Contents Pregnancy diagnosis by palpation of rectum is routine

The incidence of early foetal loss is increasing under intensive practice in reproductive management programmes, and
management systems for dairy cattle. The aims of the present it is generally performed just before the foetal period,
study were to determine whether there is any peak period of between 30 and 40 days of gestation. Early foetal loss is
pregnancy loss during the early foetal period and to evaluate assumed to be of multi-factorial origin and often escapes
possible differences between single and twin pregnancies. The clinical diagnosis, with the risk of increase in intensive
study population consisted of 1442 pregnant cattle from a management systems (Forar et al. 1995). Indeed, foetal
single herd. Pregnancy was diagnosed by transrectal ultraso-
loss per diagnosed pregnancy can exceed 12% (Day
nography between 36 and 42 days after insemination, and then
weekly until day 90 of gestation or until pregnancy loss. A et al. 1995; Markusfeld-Nir 1997). In previous studies
total of 1310 cows (90.8%) bore single embryos and 132 (López-Gatius et al. 1996, 2002, 2004; Labèrnia et al.
(9.2%) carried twins. Pregnancy loss was registered in 139 1996), we evaluated the effects of management and cow
(9.6%) cows before day 90 of pregnancy: 101 (7.7%) in single factors on pregnancy loss following pregnancy diagno-
and 38 (28.8%) in twin pregnancies. The average time of sis. Most of the pregnancy losses were registered
pregnancy loss for all animals was 58.4 ± 12.6 days and between 38 and 90 days of gestation, and the risk of
ranged from 45 to 90 days. Seventy-five per cent of the early foetal loss was 3.1 times greater for cows bearing
pregnancy losses were registered between 45 and 60 days of twins than for those with singletons. The aims of the
gestation. The average time of pregnancy loss for cows with present study were to determine whether there is any
singletons was 52.1 ± 4.1 days and ranged from 45 to 61 days
peak period of pregnancy loss during the early foetal
and that for those with twins was 75.1 ± 12.4 days and
ranged from 46 to 90 days. Seventy-five per cent of the twin period (between 43 and 90 days of gestation) and to
pregnancy losses were registered between 68 and 90 days of evaluate possible differences between single and twin
gestation. Our data show that the foetal loss in singleton pregnancies. The effects of cow and management var-
pregnancies occurs earlier than in twin pregnancies. Assess- iables previously found to be significantly correlated
ment of normal development of gestation on days 60 and 90 with early foetal loss in the same area were also analysed
after insemination is suggested. for both single and twin pregnancies.

Introduction Materials and Methods

Although genetic progress and improvements in nutri- Animals and artificial insemination
tion and management practices have to led to a This study was performed on a commercial dairy herd of
continuous increase in milk yield over the last decades, 250 (1997) to 380 (2001) mature Holstein–Friesian cows
the fertility of dairy cows has not improved substan- in north-east Spain over a 5-year period (1 March 1997
tially. On the contrary, loss of reproductive efficiency to 28 February 2002). The mean annual milk production
appears endless in high-producing dairy cattle (Lucy of the herd for this period ranged from 9200 (1997) to
2001). Numerous reports in recent years have assessed 10600 (2001) kg per cow. The cows were milked three
the simultaneous increase of milk production, repro- times daily and kept in open stalls. All animals were
ductive disorders and infertility (Opsomer et al. 2000; reared within the herd, and they were non-vaccinated,
Royal et al. 2000; Mayne et al. 2002; López-Gatius tuberculosis- and brucellosis-free, and bovine viral
2003; Rajala-Schultz and Frazer 2003). Primary atten- diarrhoea- (BVD-) and Neospora-seronegative. The
tion is usually directed towards post-partum reproduc- absence of these diseases was confirmed by a planned
tive disorders and the failure in correctly detecting monitoring programme.
oestrus. However, once a cow is pregnant, it is also of The herd was maintained on a weekly reproduc-
great interest to account for the influence of pregnancy tive health programme and the reproductive tract of
loss on the overall reproductive cycle. Pregnancy failure each animal examined by palpation of rectum within
is probably the single most economically important 30–36 days post-partum to check for normal uterine
problem faced by the dairy industry, with a greater effect involution and ovarian structures. Reproductive disor-
if losses occur between 30 and 60 days of gestation, ders diagnosed at this time were treated until resolved.
during the late embryo and early foetal periods (Ball Incomplete involution, endometritis-pyometra and
1997; Vanroose et al. 2000). The embryonic period of inactive ovaries or ovarian cysts (these two latter diag-
gestation extends from conception to the end of the nosed in a subsequent examination 7 days after) were
differentiation stage (about 42 days), and the foetal considered reproductive disorders. The selection of
period extends from day 42 to parturition (Committee the herd was based on analysis of its reproductive
on Bovine Reproductive Nomenclature 1972). history, reproductive health programme and on data

U.S. Copyright Clearance Centre Code Statement: 0936–6768/2004/3906–0429$15.000/0 www.blackwell-synergy.com

430 F López-Gatius, P Santolaria, JL Yániz, JM Garbayo and RHF Hunter

corresponding to the 2 years prior to the study, which embryos were recorded as pregnancies with an addi-
showed that following pregnancy diagnosis (performed tional corpus luteum.
between 34 and 40 days of gestation) no abortions were All ultrasonographic examinations and manual preg-
registered in heifers and that, in mature cows, 90% of nancy diagnoses were performed by the same operator.
pregnancy loss had occurred before gestation day 90.
This value of 90% before day 90 is common among
several herds with similar health management schemes, Data collection and analysis
and the herd was selected as representative. The following data were recorded for each animal:
Only cows free of detectable reproductive disorders repeated animal (animals included two or more times
were inseminated. Artificial insemination was the form within the study in which data were obtained from
of breeding used in this herd. Cows were inspected for different lactational periods), date of conception, pres-
signs of oestrus in the early morning and late afternoon ence of an additional corpus luteum, number of
for 20–30 min. Standing to be mounted was considered embryos, number of services, the interval from calving
to be the first sign of oestrus. If a registered pregnant to pregnancy, lactation number, milk production at
cow showed oestrous behaviour, it was not inseminated pregnancy diagnosis, semen-providing bull, and date of
and pregnancy was confirmed at the subsequent ultra- pregnancy loss for each animal. In the geographical area
sound examination. Thirteen animals with clinical of study, there are only two clearly differentiated
disease between first pregnancy diagnosis and 120 days meteorological periods: warm (May to September) and
of gestation were removed from the study. cool (October to April) periods (Labèrnia et al. 1998).
Reproductive parameters are generally significantly
impaired in the warm period (López-Gatius 2003) and
Pregnancy diagnosis season of insemination and pregnancy loss have been
Pregnancy was diagnosed by transrectal ultrasonogra- shown to be significantly correlated (Labèrnia et al.
phy between 36 and 42 days after insemination using a 1996; López-Gatius et al. 2002, 2004). For this reason,
portable B-mode ultrasound scanner (Scanner 100 Vet insemination dates were used to analyse the effect of the
equipped with a 5.0 MHz transducer; Pie Medical, season of insemination (warm vs cool period) on the
Maastricht, The Netherlands). Animals found to be occurrence of pregnancy loss.
pregnant were subjected to weekly (always 7 days apart
between two consecutive examinations) ultrasound
examination until day 90 of gestation or until pregnancy Timing of early foetal loss
loss. Pregnancy loss was recorded when a subsequent As a result of the clear pregnancy loss peak, and the
examination to the first pregnancy diagnosis proved evident differences between single ant twin pregnancies,
negative. The date of the negative diagnosis was descriptive statistics was only used.
registered as the date of the pregnancy loss. A further
diagnosis by palpation of rectum was also performed to
ensure a normal increase in uterine size and membrane Factors affecting early foetal loss
slip between 120 and 127 days after insemination. A cow The relative contribution of each factor to the probab-
was included only once per lactation: only the first ility of pregnancy loss in single or twin pregnancies was
diagnosed pregnancy within a lactation period was determined by logistic regression. Pregnancy loss in
entered in the study. single or twin pregnancies was considered as the
dependent variable and repeated animal, an additional
corpus luteum and season as dichotomous variables
Number of embryos (where 1 denotes presence and 0 denotes absence). Milk
Scanning was performed along the dorsal/lateral surface production, number of services, the interval from
of each uterine horn. The presence of twins was calving to pregnancy, and lactation number (continuous
established by observation of one embryo in each variables), and the semen-providing bull (class variable)
uterine horn, two embryos simultaneously present on were considered factors in the analysis.
screen scan or two embryos in different positions within Regression analysis (SAS software, logistic procedure;
one uterine horn on two screen scans (in this case, SAS 2001) was performed according to the method of
repeated observations were performed to assess the Hosmer and Lemeshow (1989). Basically, this method
presence of two separate embryos, not one and the same consists of five steps as follows: preliminary screening of
from two directions). all variables for univariate associations; construction of
a full model, using all the significant variables resulting
from the univariate analysis; stepwise removal of non-
Corpora lutea significant variables from the full model and comparison
Each ovary was scanned in several planes by moving the of the reduced model with the previous model for model
transducer along its surface to identify luteal structures, fit and confounding; evaluation of interactions among
and the number of corpora lutea was registered for each variables; and assessment of model fit using Hosmer–
cow. A corpus luteum with or without a cavity was Lemeshow statistics. Variables with univariate associa-
identified by its size and shape as well as by a granular, tions showing p-values <0.25 were included in the
grey, structured area that could be identified in the initial model. Modelling was continued until all the
ovarian tissue (Kähn 1994). Pregnancies in which the main effects or interaction terms were significant
number of corpora lutea exceeded the number of according to the Wald statistic at p < 0.05.
Early Foetal Loss in Dairy Cattle 431

The regression coefficients from the logistic regression (a)

were exponentiated to obtain the odds ratio and 60

Pregnancy losses
corresponding 95% confidence interval associated with 50
each factor. An odds ratio significantly higher than 1 40
indicates an increased (or reduced) risk of pregnancy 30
loss if the factor is present and is a dichotomous 20
variable. For continuous variables, an odds ratio 10
significantly higher (or lower) than 1 implies an 0
increased (or reduced) risk of pregnancy loss with each 43 – 48 49 – 54 55 – 60 61 – 66 67 – 72 73 – 78 79 – 84 85 – 90
1 unit increase in the value of this factor. For the class
variable, one class was considered as the reference, and (b)
an odds ratio significantly higher (or lower) than 1 for 60

Pregnancy losses
any other class of this variable was indicative of an 50
increased (or reduced) risk of pregnancy loss when 40
compared with the reference class. 30
20
10
The study population 0
The time (mean ± SD) from calving to pregnancy, 43 – 48 49 – 54 55 – 60 61 – 66 67 – 72 73 – 78 79 – 84 85 – 90
lactation number, number of inseminations and milk (c)
production at pregnancy diagnosis were 125 ± 71 days, 60
2.2 ± 1.5, 3 ± 2.2, and 33.6 ± 9.1 kg, respectively. Pregnancy losses 50
The total number of animals used was 766: 387, 196, 99, 40
54 and 30 cows were used one, two, three, four and five 30
times, respectively. Pregnancy loss for cows included in
20
the study one time (data obtained from a single lactational
10
period), or two or more times (data obtained from two or
more lactational periods) was 10.6% (41/387 pregnan- 0
43 – 48 49 – 54 55 – 60 61 – 66 67 – 72 73 – 78 79 – 84 85 – 90
cies) and 9.3% (98/1055 pregnancies), respectively. Fif- Days of pregnancy
teen cows suffered pregnancy loss twice along the study.
Pregnancy loss was registered in 9.1% (54/595) primip- Fig. 1. Timing of early foetal loss for all [n ¼ 1442 (a)], single [n ¼
arous and in 10% (85/847) multiparous cows. 1310 (b)] and twin [n ¼ 132; (c)] pregnancies
The final study population consisted of 1442 pregnant
cows bearing 1574 embryos. A total of 1310 pregnant
cows (90.8%) bore single embryos and 132 (9.2%) twins. insemination number on pregnancy loss. Based on the
No triplets were recorded. Pregnancy loss was registered odds ratios, the pregnancy loss rate was 3.7 times higher in
in 139 (9.6%) cows before day 90 of pregnancy: 101 animals that became pregnant in the warm than in the
(7.7%) in single and 38 (28.8%) in twin pregnancies. cool period, and 2.1 times higher in cows inseminated by
Two further cows with singletons were negative to one of the 13 bulls. Pregnancies with an additional corpus
pregnancy diagnosis on day 120 post-insemination and luteum were 7.7 (1/0.13) times less likely to miscarry than
were not included in the study. One hundred and seven pregnancies with no additional corpora lutea.
(7.4%) cows had an additional corpus luteum: 95 (7.3%) Season was the only variable included in the logistic
in single and 12 (9.1%) in twin pregnancies. regression model for twin pregnancies: the risk of
pregnancy loss was 5.4 times higher in animals that
became pregnant during the warm period.
Results
Timing of early foetal loss
Timing of early foetal loss for pregnancies is shown in Discussion
Fig. 1. Seventy-five per cent of the all pregnancy losses The most noteworthy finding in the present study was
were registered between 45 and 60 days of gestation. For that, during the examination period from days 43 to 120
cows bearing singletons, all pregnancy losses were of gestation, 75% of pregnancy loss was recorded
registered between 45 and 61 days of gestation. Sev- between days 45 and 60 of gestation, with the peak of
enty-five per cent of the twin pregnancy losses were loss in singleton pregnancies occurring conspicuously
registered between 68 and 90 days of gestation. earlier than in twin pregnancies. Precise reasons for the
loss were not revealed here, but could be associated with
the metabolic demands placed upon high-yielding dairy
Factors affecting early foetal loss cows at a time when placental and foetal development
Table 1 shows pregnancy loss in single and twin preg- are competing with the major drain of lactation. Such
nancies for variables selected by the logistic procedure. metabolic stress would find expression in an endocrine
In pregnancies bearing singletons, logistic regression imbalance, seemingly compromising function of the
analysis indicated no significant effects of repeated corpus luteum and its secretion of progesterone. Sup-
animals, lactation number, days in milk at pregnancy, port for this assertion is that single pregnancies with an
milk production at pregnancy diagnosis, and additional corpus luteum were eight times less likely to
432 F López-Gatius, P Santolaria, JL Yániz, JM Garbayo and RHF Hunter

Table 1. Pregnancy loss in single

Factor Class n % Odds ratio 95% confidence interval p-value and twin pregnancies for variables
selected by the logistic procedure
Pregnancy loss for pregnancies with singletons (n ¼ 1310)a
Season Cool 37/805 4.6
Warm 64/505 12.7 3.7 2.3–6 <0.0001
Additional corpus luteum 0 100/1215 8.2
1 1/95 1.1 0.13 0.02–0.97 0.046
Bull 13 classes 0.03
Bull 10 2.1 1.2–5.3 0.003
Pregnancy loss for pregnancies with twins (n ¼ 132)b
Season Cool 16/91 17.6
Warm 22/41 53.7 5.4 2.4–12.3 <0.0001

a
Likelihood ratio test ¼ 649.3; d.f. ¼ 14, p < 0.0001.
Hosmer and Lemeshow goodness-of-fit test ¼ 1.42; d.f. ¼ 7, p ¼ 0.99 (the model fits).
b
Likelihood ratio test ¼ 141.2, d.f. ¼ 1, p < 0.0001.

undergo pregnancy loss than pregnancies with no risk of pregnancy loss was 5.7 times higher for cows that
additional corpus luteum in agreement with a previous became pregnant during the warm period.
experiment (López-Gatius et al. 2002), and progesterone Data presented here show a total incidence of foetal
supplementation has the potential to reduce the inci- loss in 9.6% of pregnancies. Ball (1997) suggested that
dence of pregnancy loss during the early foetal period 5% of pregnancies ending in late embryo/early foetal
(López-Gatius et al. 2004). Furthermore, it is accepted death may be a basic level of loss, and losses in excess of
that progesterone plays a major role in controlling the these levels are likely because of specific causes. In our
maternal secretion of nutrients, growth factors and study region, warm season is a main cause of early foetal
immunosuppresive agents required for successful em- loss. These findings prompt the need to improve
bryo/foetal development and pregnancy recognition management practices by attempting to reduce the
(Ford 1985; Mann and Lamming 1999; Spencer and effects of heat stress during the warm period. This need
Bazer 2002). acquires special relevance in cows with twins. The figure
The development of methods to monitor trophoblas- of 53.7% (22/41) of losses for cows bearing twins that
tic function, as in the measurement of placental proteins became pregnant during the warm period contrasts
named pregnancy-associated glycoprotein (PAG), preg- clearly with that 4.6% (37/805) losses for cows with
nancy-specific protein B (PSPB) or pregnancy-specific singletons that became pregnant during the cool period.
protein 60 (PSP-60) (Green et al. 1998), should help to Twin pregnancies are undesirable in dairy cattle,
elucidate whether embryo viability is compromised even because the profitability of the herd diminishes drastic-
when the embryo is still alive (Mialon et al. 1993). ally as the frequency of twin births increases (Eddy et al.
Moreover, in the case of twin pregnancies, PAG 1991), and increased milk production has been described
measurement would allow us to determine if embryo over the last two decades as the single most important
losses are really occurring later in gestation or if one cause for the concurrent increase in the rate of twins
foetus is potentially ‘protecting’ the other for a period of (Nielen et al. 1989; Kinsel et al. 1998). It can be
time. Cows carrying twins with one dead conceptus can expected that the rate of twins will increase over the
maintain gestation for several weeks before pregnancy following years, along with increased milk production.
loss (López-Gatius and Hunter 2004), and in goats with In the present study, 9.2% of total cows carried twins
multiple gestations, a drop in PAG concentration occurs and 29% of them suffered pregnancy loss, figures similar
before the expulsion or loss of a foetus (Zarrouk et al. to those previous studies (López-Gatius et al. 2002;
1999). López-Gatius and Hunter 2004).
Although the effect of cow and management factors A classical means of interpreting prenatal loss in
on pregnancy loss was analysed in previous studies placental mammals concerns the reproductive cycle.
(López-Gatius et al. 1996, 2002; Labèrnia et al. 1996, Nature appears to have imposed screening programmes
2004), the possible effects of these factors were also so that the bulk of any loss occurs early in embryonic
analysed in the present study, given that these param- development, thereby causing minimal disruption to the
eters had large ranges of variation. Our results were breeding cycle. In other words, with the majority of loss
similar to those previous studies: lactation number, days featuring as early embryo loss, and in the absence of
in milk at pregnancy, milk production at pregnancy infection, the animals can return to oestrus, undergo
diagnosis, and insemination number were not related to mating, and re-establish pregnancy with loss of time
pregnancy loss. Neither it could be related with the rarely exceeding the duration of two oestrous cycles,
pregnancy loss the inclusion of practically half of the usually significantly less. This may be essential in species
animals several times (in different lactational periods) or breeds in which seasonal patterns of breeding are
along the study. A strong association was however shown. However, one of the major setbacks to repro-
observed between warm period, bull (as negative ductive efficiency associated with selection for high
factors) and an additional corpus luteum (as a positive lactation yield, besides infertility increase, is the relat-
factor), and pregnancy loss in single pregnancies. These ively belated phase of foetal loss and, thereby, a major
results add strength to the already published data delay in the potential time of re-breeding – in fact a
(López-Gatius et al. 2002). Finally, the only factor that major shift in the reproductive cycle. It needs to be
dramatically affected twin pregnancies was season; the appreciated that the overall cost of such delay and
Early Foetal Loss in Dairy Cattle 433

modified breeding cycle may far exceed the value of the López-Gatius F, Hunter RHF, 2004: Spontaneous reduction
extra milk production for which animals were selected in of advanced twin embryos: its occurrence and clinical
the first instance. For example, it was estimated that the relevance in dairy cattle. Theriogenology in press.
cost of a pregnancy loss during the early foetal period López-Gatius F, Labèrnia J, Santolaria P, López-Béjar M,
Rutllant J, 1996: Effect of reproductive disorders previous to
amounts to US $640 (Thurmond and Picanso 1990).
conception on pregnancy attrition in dairy cows. Therioge-
Considering 10% early foetal loss, we could thus nology 46, 643–648.
calculate that the annual cost to a dairy herd with 100 López-Gatius F, Santolaria P, Yániz J, Rutllant J, López-
pregnancies per year could be as high as $6400 [(0.1) Béjar M, 2002: Factors affecting pregnancy loss from
(100) (640)]. Mathematical modelling of the variables gestation day 38 to 90 in lactating dairy cows from a single
involved should reveal the dubious advantage of further herd. Theriogenology 57, 1251–1261.
rigorous selection for only milk yield in dairy cattle. López-Gatius F, Santolaria P, Yániz J, Hunter RHF, 2004:
The risk of pregnancy loss was practically non-existent Progesterone supplementation during the early fetal period
after day 60 of gestation in cows with singletons, the reduces pregnancy loss in high-yielding dairy cattle. Theri-
period in which pregnancy is considered to be firmly ogenology in press.
Lucy MC, 2001: Reproductive loss in high-producing dairy
established and the chances of loss are reduced (Ball
cattle: where will it end? J Dairy Sci 84, 1277–1293.
1997). However, our results show that twin pregnancies Mann GE, Lamming GE, 1999: The influence of progesterone
appear to need a longer period for a firm establishment of during early pregnancy in cattle. Reprod Dom Anim 34,
pregnancy, a period which reached day 90 of gestation. 269–274.
From a practical perspective, our results suggest that Markusfeld-Nir O, 1997: Epidemiology of bovine abortions in
additional pregnancy diagnoses on days 60 and 90 after Israeli dairy herds. Prev Vet Med 31, 245–255.
insemination should be included in routine management Mayne CS, McCoy MA, Lennox SD, Mackey DR, Verner M,
programmes to prompt detect early foetal loss. Catney DC, McCaughey WJ, Wylie ARG, Kennedy BW,
Gordon FJ, 2002: Fertility of dairy cows in Northern
Ireland. Vet Rec 150, 707–713.
Acknowledgements Mialon MM, Camous S, Renand G, Martal J, Menissier F,
1993: Peripheral concentrations of a 60 kDa pregnancy
The authors would like to thank Paqui Homar for assistance
serum protein during gestation and after calving and in
with the data collection. R.H.F. Hunter was supported by a
relationship to embryonic mortality in cattle. Reprod Nutr
grant from the University of Lleida (Project No. 08121).
Dev 33, 269–282.
Nielen M, Schukken YH, Scholl DT, Wilbrink HJ, Brand A,
1989: Twinning in dairy cattle: a study of risk factors and
References effects. Theriogenology 32, 845–862.
Ball PJH, 1997: Late embryo and early fetal mortality in the Opsomer G, Grohn YT, Hertl J, Coryn M, Deluyker H, de
cow. Anim Breeding Abs 65, 167–175. Kruiff A, 2000: Risk factors for post partum ovarian
Committee on Bovine Reproductive Nomenclature, 1972: dysfunction in high producing dairy cows in Belgium: a field
Recommendations for standardizing bovine reproductive study. Theriogenology 53, 841–857.
terms. Cornell Vet 62, 216–237. Rajala-Schultz PJ, Frazer JS, 2003: Reproductive performance
Day JD, Weaver LD, Franti CE, 1995: Twin pregnancy in Ohio dairy herds in the 1990s. Anim Reprod Sci 76, 127–
diagnosis in Holstein cows: discriminatory powers and 142.
accuracy of diagnosis by transrectal palpation and outcome Royal MD, Darwash AO, Flint APF, Webb R, Woolliams JA,
of twin pregnancies. Can Vet J 36, 93–97. Lamming GE, 2000: Declining fertility in dairy cattle:
Eddy RG, Davies O, David C, 1991: An economic assessment changes in traditional and endocrine parameters of fertility.
of twin births in British dairy herds. Vet Rec 129, 526–529. Anim Sci 70, 487–501.
Forar AL, Gay JM, Hancock DD, 1995: The frequency of SAS, 2001: Technical Report: Release 8.2, 2001. SAS Institute
endemic fetal loss in dairy cattle: a review. Theriogenology Inc., Cary, NC.
43, 989–1000. Spencer TE, Bazer FW, 2002: Biology of the progesterone
Ford SP, 1985: Maternal recognition of pregnancy. Therioge- action during pregnancy recognition and maintenance of
nology 23, 145–159. pregnancy. Front Biosci 7, D1879–D1898.
Green JA, Xie S, Roberts RM, 1998: Pepsin-related molecules Thurmond MC, Picanso JP, 1990: A surveillance system for
secreted by trophoblast. Rev Reprod 3, 62–69. bovine abortion. Prev Vet Med 8, 41–53.
Hosmer DW, Lemeshow S, 1989. Applied Logistic Regression. Vanroose G, de Kruif A, Van Soom A, 2000: Embryonic
Wiley, New York. mortality–pathogen interactions. Anim Reprod Sci 60–61,
Kähn W, 1994: Veterinary Reproductive Ultrasonography. 131–143.
Schlütersche Verlagsanstalt und Druckerei GmbH & Co, Zarrouk A, Engeland I, Sulon J, Beckers JF, 1999: Determin-
Hannover, pp. 83–210. ation of pregnancy-associated glycoprotein concentrations
Kinsel ML, Marsh WE, Ruegg PL, Etherington WG, 1998: Risk in goats (Capra hircus) with unsuccessful pregnancies: a
factors for twinning in dairy cows. J Dairy Sci 81, 989–993. retrospective study. Theriogenology 51, 1321–1331.
Labèrnia J, López-Gatius F, Santolaria P, López-Béjar M,
Rutllant J, 1996: Influence of management factors on preg- Submitted: 01.06.2004
nancy attrition in dairy cattle. Theriogenology 45, 1247–1253.
Labèrnia J, López-Gatius F, Santolaria P, Hanzen C, Laurent Author’s address (for correspondence): F López-Gatius, Departamen-
Y, Houtain JY, 1998: Influence of calving season on the to de Producción Animal, Universidad de Lleida, Escuela Técnica
interactions among reproductive disorders of dairy cows. Superior de Ingenierı́a Agraria, Avda. Alcalde Rovira Roure 177,
Animal Sci 67, 387–393. 25198 Lleida, Spain. E-mail: flopez@prodan.udl.es
López-Gatius F, 2003: Is fertility declining in dairy cattle? A
Present address: Department of Clinical Studies – Reproduction,
retrospective study in northeastern Spain. Theriogenology Royal Veterinary University, Copenhagen, Denmark
60, 89–99.