Management of summer infertility in Texas Holstein dairy cattle

ELSEVIER

MANAGEMENT

OF SUMMER INFERTILITY

IN TEXAS HOLSTEIN DAIRY CATTI-E

J.A. Thompson, la D.D. Magee,’ M.A. Tomaszewski,*

D.L. WilksI and R.H. Fourdraine*

‘Department of Large Animal Medicine and Surgery 2Department of Animal Science Texas A&M University College Station, TX 77843, USA Received for publication: Accepted:

December IO, 199s March. I, 19%

ABSTRACT The objectives of this study were to describe the impact of season on pregnancy odds and the effect of specific herd management practices to modify seasonal effects. Pregnancy odds were significantly associated with herd, season, days in milk, and milk production, and with 3 interactions: milk production-by-days in milk, milk production-by-parity, and season-by-days in milk. The estimate of relative risk for summer insemination resulting in pregnancy was 0.66 at 60 d post calving and 0.53 at 160 d post calving. Shade in the lounging area, holding pen or dry cow areas, and fans in the lounging area had positive effects on summer pregnancy odds. Fans Sprinklers did not in the dry cow area were associated with a reduced odds of pregnancy. significantly modify the effect of season on pregnancy odds. The strong seasonal decrease in pregnancy odds was less severe on farms that provided shade in the lounging areas, holding pens and dry cow areas and fans in lounging areas. Insemination strategies can also be adapted to increase the pregnancy odds. Key words: infertility, heat stress, management

INTRODUCTION In subtropical climates, high ambient temperature is an important determinant of reproductive performance (13). Heat stress decreases the intensity and duration of estrus, which in turn reduces both the total number of inseminations and the proportion of inseminations that result in pregnancy (13). Heat stress alters concentrations of circulating hormones by increasing circulating concentrations of corticosteroids (20) and by reducing progesterone concentration (14,22). The viability of pre-fixation embryos is reduced (5), and the uterine environment is altered by decreased blood flow (21) and increased uterine temperature (11). These changes are

Acknowledgements Support was provided through the Texas Dairy Herd Improvement aCorrespondence and reprint requests.

Theriogenology 46547-558. 1996 8 1996 by Elsevier Science Inc.

Association.

0093-691 X/96/$1 5.00 PII SOO93-691X(96)00176-3

Theriogenology

548 associated with increased inseminations.

early embryonic

loss and a reduced

proportion

of successful

Several management techniques, based on experimental evidence have been recommended to control seasonal infertility. Recommendations include the use of sprinklers and evaporative cooling, fans, shade, and their combinations (8,20) The positive effects of inseminating younger, lower producing cows during periods of heat stress have been shown, and this practice is promoted (11) The findings from experimental and small observation studies have not been confirmed by a field study involving more than a few herds. A recent, large field study reported that fertility was not as depressed as previously reported, and Ray et al. speculated that the management of seasonal effects had improved (19). The first objective of this study was to describe the impact of season on the pregnancy odds in a population of Texas dairy cattle, which consisted of Texas Dairy Herd Improvement Association (DHIA) dairy herds in the 8 counties (Erath, Hopkins, Comanche, Jotion, Wood, Archer, Cherokee and Wise), making up the major milk producing area in Texas. The second objective was to describe the value of herd management practices, specifically the effects of shades, fans and sprinklers in various locations on fertility.

MATERIALS AND METHODS Texas DHIA personnel examined the reproductive records of milk herds and provided a list of 141 herds in the 8-county milk district for which the producers selected the DHIA option to have pregnancy determinations routinely reported. Producers were contacted for permission to use their data and were asked to describe the location of sprinklers, shades and fans at their facilities. Data from participating herd managers included herd identification and cow information on breed, calving date, days to first insemination, days open, parity, and test-day milk production. When the records were obtained in February 1994, the latest month with pregnancy testing results was October 1993. Records from November 1992 to October 1993 were selected, limiting the study to a l-year period to minimize the cows for which samples would be collected twice. Initial analyses were performed using graph methods. To study seasonal effects, the log odds of pregnancy were determined by month and plotted. The log odds of pregnancy by days in milk at first insemination were calculated by treating the relationship as linear, and the resulting regression line was plotted against the actual log odds of pregnancy for 10-d intervals. Test day milk production, recorded closest to the date of insemination, was divided into 6 classes: <20, 20 to 29, 30 to 39, 40 to 49, 50 to 59 and 260 kg per day, and the log odds of pregnancy were calculated for each class and plotted. The log odds of pregnancy were calculated and plotted for Parities 1 to 4, while Parities 5 and higher were combined. Multiple logistic regression was used to model the outcome of pregnancy to the first service insemination. The herd was treated as a fixed effect by creating 118 dummy or design variables to code for 119 herds. Days in milk was coded as a linear variable. The months of July through September were defined as the summer season. Parity was divided into 2 classes, first

Theriogenology

549

lactation and later lactations. Milk production was also divided into two classes, higher production (> 30 kg/d) and lower production (C 30 kg/d). The full model was fitted and then expanded by selecting two-way interactions, using forward stepwise selection from all possible two-way interactions except for the herd interactions. The statistic used for entry was the likelihood ratio statistic, and an entry significance of P < 0.05 was used. To test the effect of modification of season, the model was expanded to include interaction variables of the 11 herd management factors with season. Variables were created for shade and fans at each of 4 locations: 1) the lounging area (pasture, dry lot and free stalls); 2) the feed alley; 3) the holding pen; and 4) the dry cow area. Variables were created for sprinklers for each of 3 locations: 1) the lounging area, 2) the feed alley and 3) the holding pen. The odds ratios and 95 % confidence intervals were calculated for each method at each location. The analyses were performed by SAS for OS12, version 6.08.

RESULTS Of the 141 identified producers, 119 (84.4%) agreed to participate. The study involved 30,358 first-service inseminations for which pregnancy was confirmed by palpation, and nonpregnancy by return to estrus or by palpation. Holsteins constituted the major breed, with 29,540 observations (97.3%). Jerseys contributed 370 observations (1.2%) and were clustered within 9 herds, all of which had at least some Holsteins. The remaining cows (1.5 X) were of less common or mixed breeds. Further modeling was restricted to Holsteins. Graphical methods demonstrated trends in simple associations with the odds of pregnancy (the ratio of the pregnant to not pregnant proportions) using the log scale for the odds of pregnancy. Test-day milk was linearly related to the log odds of pregnancy (Figure 1). Graphical depiction of the log odds of pregnancy with month showed reduced odds of pregnancy in the months of July, August and September (Figure 2). The log odds of pregnancy were higher for first parity cows than for all other parities (Figure 3). Increased milk production was associated with a reduction in the log odds of pregnancy except for the highest production group (Figure 4). In the multivariate analysis, the log odds of pregnancy remained significantly associated with herd (P < O.OOOOS),season (P < O.OOOOS),days in milk (P < 0.00005) and milk production (P < O.oooOS), and with 3 interactions: milk production by days in milk (P < O.OOOS),season by days in milk (P < 0.01) and milk production-by-parity (P < 0.05; Table 1). The interactions can best be illustrated by comparing odds ratios for groups defined by the levels of the interacting variables and comparing these to a baseline group. The highest pregnancy odds were observed for low producing, first lactation cows, in which insemination occurred later in their lactation (160 d) in seasons other than summer. All the other groups had lower pregnancy odds (Table 2). The effect of season varied with days in milk. At 60 d post calving, the effect of summer on insemination was a reduction in the odds of pregnancy by a factor of 0.53 (odds ratio = 0.53). Later in the lactation, at 160 d post calving, the odds of pregnancy were reduced by 0.66 (odds ratio = 0.66).

Theriogenology

550

Odds of pregnancy 5

0 bserved .

F

II

I

Om535 45

I

I

55

I

I

75 65

I

I

95 65

I

II

11

11

1

I

115 135 155 175 195

105125145165185

Days to first insemination Figure 1.

The odds of pregnancy plotted against days in milk at the first insemination, demonstrating the linear relationship on the log scale.

0 dds of pregnancy

Dec92

Feb 93 Apr 93 Jun 93 Aug 93 Ott 93 Month

Figure 2.

The odds of pregnancy plotted against month, demonstrating decreased pregnancy odds during June, July and August.

551

Theriogenology

0 dds of pregnancy

Parity Figure 3

The odds of pregnancy plotted against parity, demonstrating pregnancy for Parity 1.

increased odds of

Odds of pregnancy

Test day milk production (kg)

Figure 4.

The odds of pregnancy plotted against test-day milk production, demonstrating an inverse relationship that appears linear on the log scale except for the highest production group.

Theriogenology

552

Table 1.

Odds ratios for factors influencing the odds of pregnancy Odds ratio

Variable

...

Herd

95 % Confidence interval for the odds ratio ...

Season July to September vs October to June

0.54

0.38 to 0.76

Days in milk lo-day increases

1.04

1.03 to 1.05

Milk production High (> 30 kg) vs low (< 30 kg per day)

0.60

0.52 to 0.69

Parity First vs later lactations

1.00

0.91 to 1.09

Milk production-by-days in milk High producing cows and lo-day increases

1.02

1.01 to 1.03

Season-by-days in milk Summer insemination

1.02

1.01 to 1.03

1.14

1.02 to 1.28

and lo-day increases

Milk production-by-parity High milk production for the first lactation

553

Theriogenology

Table 2.

Odds ratios for groups defined by milk production, parity, days in milk and season accounting for interactions (relative to the most fertile group)

Milk

Parity

Low

First

Days in milk 160

60

Second or later

160

60

High

First

160

60

Second or later

160

60

a Arbitrarily set as baseline.

Season

Odds ratio

95 % Confidence interval for the odds ratio

October to June

1”

July to September

0.66

0.43 to 1.01

October to June

0.66

0.61 to 0.72

July to September

0.35

0.25 to 0.50

October to June

1.00

0.91 to 1.09

July to September

0.66

0.43 to 1.02

October to June

0.66

0.59 to 0.74

July to September

0.35

0.25 to 0.51

October to June

0.86

0.68 to 1.10

July to September

0.57

0.35 to 0.93

October to June

0.46

0.41 to 0.51

July to September

0.24

0.17 to 0.35

October to June

0.75

0.56 to 1.01

July to September

0.50

0.30 to 0.83

October to June

0.40

0.33 to 0.48

July to September

0.21

0.14 to 0.31

. .

Theriogenology

554

Farm management practices that modified seasonal effects included shades and fans but not sprinklers (Table 3). Shade in the lounging area ( P < O.OS), holding pen ( P < O.OOOl), and dry cow area ( P < 0.01) had positive associations with the summer pregnancy odds; shade in the feed alley was not significantly ( P = 0.15) associated with pregnancy. Fans in the lounging area ( P < 0.01) were related to increased odds of pregnancy, and fans in the dry cow area ( P < 0.001) with a decreased odds of pregnancy. Fans in the holding pen and feed alley were not significantly ( P > 0.1) associated with pregnancy nor were sprinklers at any of the locations (Table 3). Table 3.

Odds ratios for herd management,

Variable

season modifiers of the odds of pregnancy

Odds ratio

95 % Confidence interval for the odds ratio

Shade Lounging area

1.19

1.02 to 1.39

Holding pen

1.49

1.22 to 1.82

Feed alley

0.91

0.80 to 1.03

Dry cows

1.21

1.05 to 1.39

Lounging area

1.22

1.05 to 1.41

Holding pen

0.94

0.86 to 1.03

Feed alley

0.93

0.78 to 1.11

Dry cows

0.70

0.57 to 0.85

Lounging area

0.97

0.81 to 1.16

Holding pen

1.02

0.88 to 1.22

Feed alley

1.04

0.93 to 1.12

Fans

Sprinklers

DISCUSSION The current study showed a linear association between the log of the odds of pregnancy following the first insemination and days in milk over the interval between 30 and 200 days in milk. Days to first breeding increases in hot summer climates (6); this association could be due to a reduction in ovarian cyclicity, estrus detection sensitivity, or a management decision to withhold insemination. These associations indicated that days in milk was a potential confounder

Theriogenology of the effect of season on fertility. Days in milk interacted with milk production, parity and season, and these higher order variables were included in the final model to prevent their confounding of the effects of season or season modifiers. In the current study, by the time of the first insemination, first and later lactations were not significantly different in fertility when matched for low milk production (~30 kg per day). High production was associated with reduced pregnancy in both first and later lactations, but first lactation cows in the higher production group had a smaller detrimental effect. Within the higher production group there was possible residual confounding by production, because first lactation cows could have been nearer to the low end of the range than later lactation cows. The present study used 2 standards for coding milk production to adjust for this potential confounder, and this coding was sufficient to demonstrate similarity among lower producing cows. An effect of season on first lactation cows has been previously demonstrated in 6 North Carolina dairy herds (6). The interaction between parity and fertility, as mediated by production, may explain contradictory findings on fertility associated with parity (19). Previous evidence of a potential interaction was observed in a single experimental herd in which young cows had a summer pregnancy proportion of 42.7%, while older cows had a pregnancy proportion of 31.9% (12); but that earlier study failed to demonstrate a statistical significance, possibly because the study was too small to examine complex parity and production interrelationships. In the postpartum period, there is additional stress on first lactation cows to grow, produce milk and return to sound reproductive condition following the first calving as result of relative energy deficiency (3). The common manifestation of this deficiency is anestrus, which is associated with abnormal ovarian follicular development (3). The results in our present study showed no difference in pregnancy odds for low producing cows, but this fertility parameter was evaluated only on cows that were deemed to be in estrus. Management of heat stress by inseminating younger, lower producing cows has been recommended (5). This management tool might be useful if heifer replacements are inseminated to calve just prior to spring in order to schedule summer insemination dates. With advancing parity (and production), insemination dates will tend to advance into autumn and winter on subsequent parities. The alternative strategy of withholding summer insemination in mature and high producing cows could improve pregnancy odds following insemination but may constitute unsound economic practice (4). However, withholding insemination might be economically sound when using expensive semen for cows with a very low likelihood of becoming pregnant. The odds of pregnancy reported here could be used to determine optimal insemination periods. High milk production increased the number of days open by means of 2 effects: increased days to first The current study adjusted the breeding and decreased first service pregnancy proportion. estimate of season for potential confounding by milk production and its various interactions. Season had a highly significant statistical and biological effect that was modified by days in milk. At 60 d in milk the odds of pregnancy were reduced by 0.53 times (odds ratio = 0.53) for summer insemination, and at 160 d in milk by 0.66 times (odds ratio = 0.66). Both odds ratios demonstrated clinically important decreases in the pregnancy odds. Heat stress has been reported to act through several biological mechanisms. Increased temperatures act on the ovaries, lowering progesterone concentrations (14,20) and altering oocytes, retarding the development of the dominant follicle and, subsequently, oocyte function (2,15). Motor activity and manifestations

556

Theriogenology

of estrus are reduced (13), leading to reduced accuracy of heat detection accuracy and to 2 counter-acting related effects: improved pregnancy odds associated with increased days in milk, and reduced pregnancy odds because of the difficulties in timing insemination (9). High ambient temperature affects the preattachment stage embryo (18) but with decreasing effect as the embryo develops (5). Effects on the uterus include alterations in the secretion of calcium and protein (10,16). The interaction of season with days in milk suggests that the 2 risk factors may share a common biological mechanism (17). In the current study, however, we could not determine the specific mechanism because endocrine function, uterine function and oocyte quality are all associated with days in milk (3,7) and are also proposed as mechanisms for the effects of season (13). Identification of this interaction, however, suggests that improving fertility in the early postpartum period with nutritional management (3,7) and disease prevention (24) would have a direct effect on pregnancy odds but would also modify the effect of season. Of the management factors studied, shade provided the most consistent alleviation of the effects of summer temperatures, as has been demonstrated in a Florida study (20); however, not ah locations provided the same effect. Shade was effective in lounging areas (free stalls, pasture and dry lot) and in the holding pen and with the dry cows. Shade in the feed alley was not associated with the pregnancy odds, possibly because additional comfort at the feeding alley might be associated with factors that mediate a reduced pregnancy proportion such as changes in the detection of estrus. Fans were effective if positioned in lounging areas where cows spend the most time during hot weather (8). For the dry cows, fans were associated with decreased odds of pregnancy following calving. This finding confirms an earlier one by Wolfenson et al. (25) that prepartum cooling reduces the first month milk production in cows that calve in early summer. Wolfenson et al. (25) hypothesized that the prepartum cooling may have delayed heat adaptation in this group of cows. Our current study showed that shade and fans in the dry cow area had opposite effects, with shade being beneficial and fans detrimental to subsequent pregnancy odds. The use of fans may possibly be associated with the closer confinement and reduced exercise or feed intake in the summer months reIative to that of dry cows provided with shade. Sprinklers were not found to modify the effects of season on pregnancy odds, in contrast to a more rigidly controlled study (8). It is possible that the experimental strategy was not followed adequately in the study farms or, alternatively, the experimental conditions were altered. It is possible that the relative humidity in Texas may be too high for evaporative cooling. For the month of July, the 30-year mean day-time high in Dallas/Fort Worth was 35.8 C and the mean morning relative humidity was 81% (1). Estimation of the odds ratios on cow-level factors was based on a non-random sample which was selected for the quality of reproductive data. This sampling was performed to protect the study’s internal validity at the expense of the ability to apply the results to other populations (23). Although the current study presents sound information on the magnitude of seasonal effects in the study population, it may not be applicable to other geographical regions. Infertility was

557

Theriogenology

also attributed to parity and days in milk and various interactions. Other than season, the cowlevel risk factors reasonably apply to a wide range of management systems and locations. The introduction of farm management variables (fans, shades, sprinklers) represented the introduction of ecological factors. We cannot be certain which cows spent time near the fans, shades and sprinklers and which cows did not. We only know that the devices were available to the population. An ecological study is not the best design for an estimation of effects but it appropriately tested how these farm management systems are currently operating in the study population. In the study population of Texas cows, there was a strong seasonal decrease in pregnancy odds that was less severe on farms using shade in lounging areas and holding pens and in the dry cow areas as well as fans in lounging areas. Dairy producers could also reduce summer infertility by inseminating cows with lower milk production and by incorporating nutritional practices and preventive medicine programs that increase fertility early in the postpartum period.

REFERENCES 1. 2.

3. 4. 5. 6. 7. 8.

9.

10. 11.

12.

Anon. Local Climatological Data. Ashville: National Climatic Data Center, 1994;3 Badinga L, Thatcher WW, Diaz T, Drost M, Wolfenson D. Effect of environmental heat stress on follicular development and steroidogenesis in lactating Holstein cows. Theriogenology 1993; 39:797-810. Britt JH. Follicular development and fertility: potential impacts of negative energy balance. Proc Nat Reprod Symp 1994; 103-l 12. Dijkhuizen AA, Stelwagen J, Renkema JA. Economic aspects of reproductive failure in dairy cattle. I. Financial loss at farm level. Prev Vet Med 1985; 3:251-263. Ealy AD, Drost M, Hansen PJ. Developmental changes in embryonic resistance to adverse effects of maternal heat stress in cows. J Dairy Sci 1993; 76:2899-2905. Faust MA, McDaniel BT, Robison OW, Britt JH. Environmental and yield effects on reproduction in primiparous Holsteins. J Dairy Sci 1988; 71:3092-3099. Ferguson JD. Patterns in fertility in dairy cows. Proc Nat Reprod Symp 1994; 113-120. Flamenbaum I, Wolfenson D, Mamen M, Berman A. Cooling dairy cattle by a combination of sprinkling and forced ventilation and its implementation in the shelter system. J Dairy Sci 1986; 69:3140-3147. Gaines JD, Thomas CB, Eicker S, Rhoda D, Goedtel M, Larson PL, Lindner ER, Lindborg M. Interestrous interval profiles of 71 Wisconsin dairy herds and relationship to days to pregnancy. Prev Vet Med 1992; 12:47-58. Geisert RD, Zavy MT, Biggers BG. Effect of heat stress on conceptus and uterine secretion in the bovine. Theriogenology 1988; 29: 1075-1082. Gwazdauskas FC, Thatcher WW, Wilcox CJ. Physiological, environmental, and hormonal factors at insemination which may affect conception. J Dairy Sci 1973; 56:873-877. Gwazdauskas FC, Wilcox CJ, Thatcher WW. Environmental and managemental factors affecting conception rate in a subtropical climate. J Dairy Sci 1975; 58:88-92.

558

Theriogenology

13.

Hansen PJ. Causes and possible solutions to the problem of heat stress in reproductive management of dairy cows. Proc Nat Reprod Symp 1994; 16 l- 170. Howell JL, Fuquay JW, Smith AE. Corpus luteum growth and function in lactating Holstein cows during spring and summer. J Dairy Sci 1994; 77:735-739. Lenz RW, Ball GD, Liebfried ML, Ax RL, First NL. In vitro maturation and fertilization of bovine oocytes are temperature-dependent processes. Biol Reprod 1983; 29: 173 . Malayer JR, Hansen PJ, Buhi WC. Effect of day of the oestrous cycle, side of the reproductive tract and heat shock on in vitro protein secretion by bovine endometrium. J Reprod Fertil 1988; 84:567 . Martin SW. Causal structures and statistical observations: their interrelationships. Prev Vet Med 1993; 18:65-73. Putney DJ, Thatcher WW, Drost M, Wright JM, DeLorenzo MA. Influence of environmental temperature on reproductive performance of bovine embryo donors and recipients in the southwest region of the United States. Theriogenology 1988; 30:905922. Ray DE, Halbach TJ, Armstrong DV. Season and lactation number effects on milk production and reproduction of dairy cattle in Arizona. J Dairy Sci 1992; 75:2976-2983 Roman-Ponce H, Thatcher WW, Buffrngton DE, Wilcox CJ, Van Horn HH. Physiological and production responses of dairy cattle to a shade structure in a subtropical environment. J Dairy Sci 1977; 60:424-430. Roman-Ponce H, Thatcher WW, Caton D, Barron DH, Wilcox CJ. Thermal stress effects on uterine blood flow in dairy cows. J Anim Sci 1978; 46: 175-180. Rosenberg M, Herz Z, Davidson M, Folman Y. Seasonal variations in post-parturn plasma progesterone levels and conception in primiparous and multiparous dairy cows. J Reprod Fertil 1977; 51:363-367. Rothman KJ. Modern Epidemiology. Boston: Little, Brown and Company, 1986;95-96. Shanks RD, Freeman AE, Berger PJ. Relationship of reproductive factors with interval and rate of conception. J Dairy Sci 1979; 62:74-84. Wolfenson D, Flamenbaum I, Berman A. Dry period heat stress relief effects on prepartum progesterone, calf birth weight, and milk production. J Dairy Sci 1988; 71:809-818.

14. 15. 16.

17. 18.

19. 20.

21. 22.

23. 24. 25.