Considering the production-reproduction antagonism and the increased culling rates of cows in intensive dairy operations ( Chiumia et al ., 2013 ; Berry et al ., 2016 ), newborn calf survival is critical for maintaining the profitability of dairy herds. Loss of calves at parturition also represent significant costs because this periparturient disorder often results in compromised reproductive performance, reduced milk yield and longevity of cows ( Berry et al ., 2007 ; Bicalho et al. , 2008 ).

Recent studies show that the prevalence of perinatal mortality in dairy calves is increasing, particularly in primiparous Holstein cows ( Mee et al. , 2014 ). Difficult calving is often the primary cause, but only accounts for approximately 50% of stillbirths causes ( Eriksson et al ., 2004 ; Barrier et al ., 2013 ). Thus, a substantial part of the stillbirths does not coincide with calving difficulties ( Mee et al. , 2008 ).

Stillbirth is usually multi-factorial in nature. Some risk factors that have been linked with this condition is twinning ( Silva del Rio et al. , 2007 ; Waldner, 2014 ), parity and dystocia ( Meyer et al ., 2000 ), calf gender ( Johanson & Berger, 2003 ), calf birth weight ( Johanson & Berger, 2003 ), serum hormones alteration ( Kindahl et al ., 2002 ) and calves management ( Palombi et al ., 2013 ), among others. Other congenital ( Berglund et al ., 2003 ), infectious ( Grooms, 2006 ), nutritional, body energy reserves ( Chassagne et al ., 1999 ) factors can contribute to the occurrence of stillborn calves. Additionally, abortive microorganisms such as Leptospira spp ( Parvez et al ., 2015 ), Brucella spp, Neospora caninum ( Brickell et al ., 2010 ), bovine viral diarrhea and infectious bovine rhinotracheitis ( Berglund et al ., 2003 ), could cause weak calves at calving compromising calves’ survival within the 24 h of life. However, around 35% of the time the cause of death of a newborn, during or within the first 24 h after birth, cannot be determined ( Berglund et al ., 2003 ).

Heat stress disrupts both physiological processes and reproductive performance of high-yielding dairy cows ( De Rensis et al ., 2015 ). The consequences of this thermal stress are not only manifested during periods of high ambient temperature, but present long-lasting effects on the fertility of the dairy cow ( García-Isperto et al., 2006 , 2007 ). Recent evidence shows that offspring from pre-partum heat-stressed cows have impaired cell-mediated immune function and compromised passive immunity ( Tao & Dahl, 2013 ), as well as a reduced secretion of prolactin, insulin and insulin-like growth factor ( Guo et al., 2016 ), which could be linked to the occurrence of stillbirths. A missing piece of the stillbirth puzzle, thermal stress during the peripartum period, has not been adequately investigated relative to stillbirth incidence. To address this gap in our understanding of stillbirth problems, the objective of this study was to test the hypothesis that high ambient temperature during the peripartum period would increase the rate of stillborn calves. An additional objective was to investigate birth weight, gestation length, semen characteristics (conventional or sexed), gender of calves, parity and type of parturition (normal or dystocic) as predictive indicators of stillbirth risk under heat stress for the most part of the year.

Throughout the study period, 7.3 (2138/29405; 95% CI 7.0–7.6) of every 100 calving events had a stillborn calf. This value falls within the expected incidence rate of 7% observed in other intensive dairy operations in temperate zones ( Chassagne et al ., 1999 ; Johanson & Berger, 2003 ; Kayano et al., 2016 ). In hot environments, the prevalence rates for stillbirths in Holstein cows range from 3 to 15 ( Hossein-Zadeh et al ., 2008 ; El-Tarabany, 2015 ). Significantly more cows suffering moderate heat stress (THI 73 to 83 units) during the eighth month of gestation presented stillbirths ( Table 1 , p <0.01) compared to cows not experiencing heat stress in the final stages of gestation. Additionally, stillbirths were more common if cows were exposed to moderate hyperthermia at calving than cows not experiencing heat stress at delivery. The percentage of pregnancies ending in stillbirths did not differ significantly between cows with moderate heat stress (THI 73 to 83 units) during the seventh month of gestation that in the group of cows not suffering heat stress at this stage of pregnancy ( Table 1 ). These data showed that even moderate heat stress did not allow cows to restore their thermal balance, presumably causing physiological changes that had an impact on the survival of calves at parturition or shortly after birth.

Cows experiencing severe heat stress at calving (THI > 83) were 1.3 times more likely to have stillborn calves than cows with less severe climate conditions at parturition ( Table 1 ). Likewise, cows undergoing extreme heat stress at the 7 ^th and 8 ^th month of pregnancy were more likely to deliver stillborn calves than cows without heat stress during the final months of pregnancy. These results suggest that under the climatic conditions of the present study, in utero heat stress in late gestation resulted in a greater incidence of stillbirths in Holstein cows, relative to cows that did not experience hyperthermia at the end of pregnancy. The way hot weather increases the risk of stillbirths is not quite clear, as no controlled experiments to find out this association have been carried out. Given that calves developing in the uterus of heat-stressed dams are exposed to increased maternal respiration rate and core temperature during the final weeks of pregnancy, their structural and physiological integrity may be compromised. The homeokinetic changes to regulate body temperature of cows suffering heat stress provoke a redistribution of blood ow from the body core to the periphery, which reduces perfusion of the placental blood vessels ( Hansen, 2009 ). Additionally, heat stress during gestation can compromise placental development, which results in fetal hypoxia, malnutrition, and eventually fetal growth impedance ( Tao et al ., 2012 ). Maternal hyperthermia also desensitizes a calf’s stress response and alters the fetal development by reducing the secretion of insulin-like growth factor-I, prolactin, and insulin ( Guo et al ., 2016 ). Additionally, maternal heat stress during late gestation alters blood metabolite profile and increases non-insulin dependent glucose uptake ( Monteiro et al ., 2016a ) which eventually negatively affects the survival of calves ( Monteiro et al ., 2016b ).

Thus, extreme thermal stress in uterus seems to alter the feto-maternal interaction compromising the fetus development and even can cause the demise of the fetus between 43 and 260 days of pregnancy as it has been observed by Mellado et al. (2016) .

Gender of the calf was associated with the occurrence of stillbirths, as bull calves were more ( p <0.001) likely to be stillborn than heifer calves ( Table 1 ). These results are consistent with several studies where it has been reported that male calves have higher odds of stillbirths than female calves ( Maltecca et al ., 2006 ; Dhakal et al. , 2013 ). In contrast to these results, Meyer et al. (2001) observed that female calves had 12% higher odds for stillbirth than male calves if they were from multiparous cows, but 7% lower odds if they were born from primiparous cows. Olson et al. (2009) found that Heifer calves were as likely as bull calves to be stillborn. Causes for these apparent contradictions are not clear, although these differences could be explained by the inconsistent definition of stillbirths among studies and the difference number of observations used in different studies.

Cows with gestations >278 were at decreased risk of stillborn calves than cows with pregnancies shorter than 278 d ( Table 1 ). These results closely coincide with results of Meyer et al. (2000) who observed that stillbirth rate increased from 23.8% among calves born at or above the mean gestation length to 55.3% for those calves born -15 to -12 d below the mean gestation length. However, gestation lengths above the average gestation period for Holstein cows also results in increased stillbirths which gave a curvilinear relationship between gestation length and incidence of stillbirths ( Fig. 1a ).

This tendency closely coincides with other studies with dairy calves ( Johanson & Berger, 2003 ). THI at seven and eight months of pregnancy and at parturition was not associated with gestation length (r= -0.02 to -0.04), therefore, clearly climatic conditions have no influence on the duration of pregnancy. The risk was much higher for calves born after a short gestation than a prolonged gestation, apparently due to organ immaturity ( Briana & Malamitsi-Puchner, 2013 ; Sangild et al. , 2013 ; Caminita et al ., 2015 ), apparently due to reduced exposure to cortisol. This hormone promotes maturation of fetal organ systems while diminishing the rate of proliferation ( Murphy et al ., 2006 ). This variable has been considered the next most important predictor of stillbirth after dystocia ( Meyer et al. , 2000 ).

Lighter calves had six times the odds of being stillborn than heavier calves at birth ( Table 1 ). Other studies have reported a trend for higher mortality risk when calf birth weight was small ( Tarrés et al ., 2005 ). Conversely, Linden et al . (2009) did not find an association of calf birth weight with stillbirth incidence in Holstein cows. Calves developing in the uterus of heat-stressed dams, as was the case in the present study, are unprotected from the increased core temperature during the end of gestation. This results in a negative consequence in the calf, such as lower birth weight ( Dahl et al ., 2016 ), and there is a consensus regarding the negative impact of low birth weights on calf survival ( Bleul, 2011 ). Also, gestation length is typically reduced by 4 to 5 d with hyperthermia during the dry period, which may account for smaller calf size with lower vitality ( Tao & Dahl, 2013 ). Gestation length significantly explained variation ( p <0.01) for birth weight, with an increase of 0.37 ± 0.004 kg of birth weight per day of gestation, which closely agrees with observations of Dhakal et al . (2013) . As birth body weight increased or decreased, the risk of stillbirths increased in a curvilinear manner ( Fig. 1b ). From <34 kg of birth weight, there was a strong change in the increased risk of stillbirths. Then, from >46 kg, each kg increase leads to increasing risk of stillbirths.

Odds of occurrence of stillborn calves were 25 times greater for cows with dystocic parturition than cows calving unaided. This value is close to that reported by Lombard et al . (2007) with mostly Holstein cows in intensive dairy operations. In cattle, dystocia has been found to be related to several negative outcomes, such as survival of newborn calves ( Eriksson et al. , 2004 ; Bicalho et al ., 2008 ; Barrier et al ., 2013 ). This is so because dystocial calves normally experience higher physiological stress, acidosis and hypoxic state which result in lower vigor ( Barrier et al ., 2012 ; Kovács et al ., 2016 ) and consequently in reduction in the transfer of passive immunity, as well as important metabolic changes ( Barrier et al ., 2013 ; Vannucchi et al ., 2015 ). Stillborn calves from dystocic parturition not necessarily exhibit specic types of trauma, but do exhibit larger lesions than their eutocic counterparts, indicating that they experienced greater trauma during delivery ( Barrier et al. , 2013 ). Nonetheless, it should be stressed that a major proportion of stillborn calves results from normal calvings ( Berglund et al ., 2003 ; Benjaminsson, 2007 ).

The use of sexed semen, season and parity of dam were not significant sources of variation for stillbirths. Of note, parity of dam has been consistently found an important risk factor for bovine stillbirths, with primiparous being at a higher risk of stillbirths than pluriparous ( Meyer et al ., 2001 ; Berry et al ., 2007 ; Bicalho et al ., 2008 ; Uematsu et al ., 2013 ). The lack of effect of this variable on stillbirths in the present study in unknown, although the causes of stillbirths vary markedly from farm to farm.

There was a distinctive increase in stillbirth rates when parturitions occurred between 18:00 and 19:00 h, compared with all other hours of the day ( Fig. 2 ). The factors responsible for this trend are unclear. One possible explanation could be just a lower attention of cows calving at this time of the day. Insufficient monitoring around parturition has a negative effect on perinatal mortality ( Gundelach et al. , 2009 ). Given that the results of the present study point toward a connection between high ambient temperature and a higher occurrence of stillbirths, it could be that the accumulated heat load throughout the day by cows about to give birth, predisposes these animals to give birth to stillborn calves with parturitions at sunset.

As conclusions, the occurrence of stillbirths on this dairy farm was associated with either moderate or extreme heat stress during the peripartum period, with negative impacts from the seventh month of pregnancy until calving. Thus, ameliorating heat stress for heifers and cows before parturition could reduce stillbirth incidence. Dystocia was another important variable affecting of whether a calf would be alive or stillborn, therefore, intervention to reduce difficult births should reduce the incidence of this problem. Both extreme birth body weight and gestation length increased the rates of stillbirths, which suggests that pre-term calves should get a special look at birth. The herd manager of this dairy operation should pay greater attention to cows giving birth at sunset because at this time of the day stillborn calves are more common.