- Email: [email protected]
Postpartum endometrial cytology in beef cows
Available online at www.sciencedirect.com
Theriogenology 71 (2009) 739–745 www.theriojournal.com
Postpartum endometrial cytology in beef cows N.R. Santos a,1, G.C. Lamb b,2, D.R. Brown b, R.O. Gilbert a,* a
Department of Clinical Sciences, College of Veterinary Medicine, Cornell University, Ithaca, NY 14850, USA b North Central Research and Outreach Center, University of Minnesota, Grand Rapids, MN 55744, USA Received 12 May 2006; received in revised form 28 August 2008; accepted 5 September 2008
Abstract The objectives were to characterize postpartum endometrial cytology and to determine the prevalence of subclinical endometrial inflammation and its impact on reproduction in beef cows. Samples for endometrial cytology (low-volume uterine lavage) were obtained from 135 of 137 Angus cows (2–87 d postpartum) in northern Minnesota, 26 d before breeding started. Agreement between examiners for subjective inflammation scores was very high (kappa = 0.971); the correlation between these scores and PMN counts was high (r = 0.83; P < 0.001), validating subjective categorization. The proportion of PMN and large mononuclear cells (principally macrophages) declined with postpartum interval (P < 0.001), whereas small mononuclear cells were consistently present (and not significantly affected by postpartum interval). Pregnancy rate to fixed-time AI was 29% and overall pregnancy rate was 89%. There was no association between cell type and ultimate pregnancy status or day of conception (P > 0.10). Although inflammation later in the postpartum period apparently impaired subsequent reproduction in dairy cows, in cows >50 d postpartum at sample collection in the present study, no cytological parameter significantly predicted final pregnancy status or day of conception. Previous twinning increased the risk of subclinical endometritis (P = 0.02), but not the probability of becoming pregnant (P = 0.14). In conclusion, we inferred that beef cows had the ability to clear uterine inflammation after resumption of ovarian cyclicity. # 2009 Elsevier Inc. All rights reserved. Keywords: Endometrial cytology; Endometritis; Postpartum uterus; Fertility; Beef cows
1. Introduction Subclinical endometritis, defined by endometrial cytology, was remarkably common in dairy cows, affecting as many as 50% of all cows 7 wk postpartum
* Corresponding author at: Department of Clinical Sciences, College of Veterinary Medicine, Cornell University, Ithaca, NY 148536401, USA. Tel.: +1 607 253 3472 fax: +1 607 253 3701. E-mail address: [email protected] (R.O. Gilbert). 1 Present address: School of Veterinary Medicine, University of Pennsylvania, New Bolton Center, 382 West Street Road, Kennett Square, PA 19348, USA. 2 Present address: University of Florida, NFREC, 3925 Highway 71, Marianna, FL 32446, USA. 0093-691X/$ – see front matter # 2009 Elsevier Inc. All rights reserved. doi:10.1016/j.theriogenology.2008.09.043
in high-producing North American dairy herds [1–5]. Furthermore, dairy cows with clinical or subclinical endometritis had severely impaired subsequent reproductive performance [1,3,6,7]. The primary impact of persistent subclinical inflammation seemed to be reduced first-service pregnancy rate, which contributed to prolonged days open. Cows with persistent subclinical endometritis were also at substantially increased risk of failure to become pregnant within 300 d and being culled [4]. Inflammatory mediators had a detrimental impact on embryo development [8]. To date, no studies have examined the prevalence or consequence of subclinical endometritis in beef cows. Given that well-managed beef cow herds typically achieve high pregnancy rates (approximately 80–90%) in
740
N.R. Santos et al. / Theriogenology 71 (2009) 739–745
a limited breeding season (60–90 d) [9], we inferred that endometritis is either uncommon in beef cows, or has limited impact on subsequent reproductive performance. Nevertheless, perhaps a small number of beef cows have subclinical endometritis that impairs subsequent reproductive performance; this would present an opportunity for strategic therapeutic intervention to increase reproductive performance by curing endometritis, or for early culling if effective treatment options were not available. Therefore, our objectives were to characterize endometrial cytology, to determine the prevalence of subclinical endometritis, and to evaluate the effect of endometritis on subsequent reproduction in a single herd of beef cattle. 2. Materials and methods 2.1. Animals This study included 137 postpartum suckled Angus cows in northern Minnesota. The age of cow, date of previous calving, degree of calving difficulty, singleton or twin birth and clinical findings at the time of sample collection were recorded. Endometrial samples for diagnosis of subclinical endometritis were obtained 26 d before the beginning of the breeding season, a day selected for convenience and to provide a range of postpartum stages for examination. Cows ranged from 2 to 87 d postpartum when endometrial samples were collected. At initiation of the breeding season, ovulation was synchronized using the CO-Synch + CIDR (Controlled Internal Drug Release) protocol [10]. Briefly, cows received a 100 mg im injection of gonadotrophin releasing hormone (OvaCyst1, Phoenix Scientific, St. Joseph, MO, USA) at the time of CIDR (EAZIBREED1, Pfizer Animal Health, New York, NY, USA) insertion and a 25 mg injection of prostaglandin F2a (Lutalyse1, Pfizer Animal Health) at CIDR removal 7 d later, followed in 48 h by a second injection of GnRH and fixed-time AI. After this initial insemination, cows were inseminated by one of the three experienced AI technicians at observed estrus during a 70-d season. Estrus was observed three times daily for at least 30 min; cows were inseminated within 12 h of detection of standing estrus. Pregnancy was diagnosed at 28–38 d after AI by transrectal ultrasonography (5-MHz intrarectal linear-array transducer, Aloka 500 V, Corometrics, Wallingford, CT, USA) and confirmed at 58– 68 d of pregnancy by a single individual. The presence or absence of pregnancy and date of successful insemination were recorded.
2.2. Sample collection and analysis Samples were obtained by aseptic infusion of 20 mL of sterile Dulbecco’s phosphate buffered saline solution via a plastic infusion catheter (MAI/Genesis 21 in flex tip pipette, Butler Animal Health Supply, Dublin, OH, USA), followed by gentle uterine agitation and aspiration of approximately 5 mL of fluid. Recovered fluid was processed within 6 h after collection by brief vortexing and centrifugation onto a glass slide in a cytocentrifuge (Cytotek cytocentrifuge, 70 g, 10 min). Slides were air dried, stained with a modified Wright-Giemsa stain (Harleco1 Hemacolor1; EMD Chemicals, Gibbstown, NJ, USA), a cover slip was attached with Cytoseal 601 (Richard-Allan Scientific, Kalamazoo, MI, USA) and slides examined under bright-field microscopy at 200 and 400. Each cow was categorized as positive or negative for endometritis by two independent examiners, using subjective assessment of the cytological preparation, as described by Gilbert et al. [4]. A subjective score of 0– 3 was assigned independently by each observer, with 0 representing the absence of inflammation and 3 severe inflammation. Independently, a third examiner counted 200 cells from each slide, categorizing them as epithelial cells, polymorphonuclear cells, large mononuclear cells (presumed to be macrophages), and small mononuclear cells (presumed to be lymphocytes). The proportion of each was calculated. All examiners were unaware of demographic information regarding each cow. 2.3. Statistical analysis The degree of agreement between examiners of the cytological preparations was determined by the calculation of the kappa statistic. The relationship between subjective assessment of endometritis and proportions of counted cells was ascertained by nonparametric analysis of variance (Kruskal–Wallis test). Pearson correlations between proportions of cell types were calculated. The effect of presence or absence and degree of inflammation, and the proportions of each cell type in cytological preparations, calving difficulty, twinning at the previous calving, age and days open at the beginning of the breeding season on the probability of becoming pregnant, and the day of the breeding season on which conception occurred was examined by multiple logistic regression and Cox’s proportional hazards regression. Differences in proportions of cows pregnant at the end of the breeding season were examined by Fishers exact test. Changes in cell populations in cytological preparations with days
N.R. Santos et al. / Theriogenology 71 (2009) 739–745
741
postpartum were examined by analysis of variance and fitted to a polynomial regression curve. Optimal cutoff points for determining the relationship between subjectively diagnosed inflammation and percentage of neutrophils in cytology samples, and for exploring the relationship between days postpartum and presence of inflammation were calculated using receiver operating characteristic (ROC) curves. Specificity, sensitivity and P-value of the ROC test are reported. Statistical calculations were performed using SigmaStat for Windows, Version 3.0 (SPSS, Inc., Chicago, IL, USA), MedCalc for Windows, version 9.6.0.0 (MedCalc Software, Mariakerke, Belgium) and StataIC 10.0 (StataCorp LP, College Station, TX, USA) software packages. 3. Results 3.1. Validation of technique Samples were obtained from 135 of 137 cows (99%); in two cases difficulty in passing the pipette through the cervix or recovering lavage fluid precluded obtaining a diagnostic sample. Weighted kappa score for agreement between examiners for assignment of subjective inflammation category was 0.971. Cows ranged from 2 to 87 d postpartum at the time of examination. Subclinical endometritis (subjectively determined) was diagnosed in 88% of cows in the first quartile (2–23 d postpartum); 34% of cows in the second quartile (24– 57 d postpartum); 19% in the third quartile (58–73 d) and 19% in the fourth quartile (P < 0.0001). The correlation between subjective inflammation score and proportion of polymorphonuclear cells in the cytological preparations was 0.83 (r2 = 0.69, P < 0.001; Fig. 1). There were differences between the groups (P < 0.001), with significant differences between Group 0 (no inflammation) and all the other groups, but not among degrees of inflammation. Based on an ROC curve, a cutoff point of >5.5% polymorphonuclear cells was highly correlated with subjective assignment as inflamed; this cutoff point was 100% specific and 78% sensitive for predicting subjectively assigned inflammation score of 1 or greater (Figs. 2 and 3). The proportions of polymorphonuclear cells and large mononuclear cells were independent of each other (r < 0.01, P > 0.9). Similarly, the proportions of large and small mononuclear cells were independent of each other (r = 0.1, P = 0.9). However, the proportions of polymorphonuclear cells and small mononuclear cells were negatively correlated (r = 0.2, P = 0.037).
Fig. 1. Percentage of polymorphonuclear cells in cytology preparations (of endometrial lavage) by subjective inflammation scores (difference among inflammation scores, P < 0.001). This figure shows median PMN%, the interquartile range, the 5th and 95th percentile, and individual outliers.
Fig. 2. Receiver Operating Characteristic (ROC) curve for percentage of polymorphonuclear cells as a predictor of subjectively assigned inflammation in beef cows. Analysis of the ROC indicated that the use of a cutoff point of more than 5.5% PMNs was 100% specific and 78% sensitive for inflammation as assessed subjectively. Area under the curve = 0.951, P = 0.0001.
742
N.R. Santos et al. / Theriogenology 71 (2009) 739–745
Fig. 3. This interactive dot diagram depicts proportion of polymorphonuclear cells in samples subjectively assessed as positive (1) or negative (0) for inflammation. Note that none of the samples judged not to be inflamed had PMN proportions greater than 5.5%. A few samples subjectively judged to be inflamed, however, had PMN proportions lower than 5.5%.
3.2. Prevalence and impact of subclinical endometritis The proportion of PMN cells declined with postpartum interval (r = 0.57, P < 0.0001). This decline most closely matched a quadratic curve (PMN% = 39 1.08 DPP + 0.008 DPP2, where DPP = days postpartum; r2 = 0.35, P < 0.0001; Fig. 4). The proportion of cows with cytologically diagnosed endometritis after 50 d postpartum was particularly low (Figs. 5 and 6). The proportion of macrophages also declined with increasing postpartum interval (P = 0.0001), but this was not true for lymphocytes (P = 0.49). Seven cows were sold after sampling but before the onset of the breeding season. None of these cows had cytological evidence of endometritis. At the time of sampling three cows were found to have severe urovagina; only one of these had evidence of endometritis. One cow had marked pneumovagina, but also lacked evidence of endometritis. All of these cows became pregnant, three on the first day of the breeding season, and the fourth 22 d after the onset of the breeding season. Pregnancy resulted from the initial insemination in 37 of 128 cows bred (29%). No cytological feature affected pregnancy to first service (P > 0.4). Pregnancy to first insemination was 28% for cows aged 2 or 3 y, 34% for those 4–6 y, and 21% for cows 7 y (P = 0.21). Days to onset of the breeding season, stratified by quartile, were examined as a potential factor affecting pregnancy to first service, but was not significant (P = 0.68).
Fig. 4. The proportion of PMN cells declined with increasing days postpartum. (PMN% = 39 1.08 DPP + 0.08 DPP2, where DPP = days postpartum; r2 = 0.35; P < 0.0001).
None of the measured factors (age, days open at onset of breeding season, calving ease, previous twinning, subjective diagnosis of endometritis, or proportion of any cell type) significantly influenced the likelihood of pregnancy at the end of the breeding season (Table 1.)
Fig. 5. Receiver Operating Characteristic (ROC) curve for the presence of inflammation by days postpartum. Using a cutoff point of 50 d postpartum predicted endometritis with 71% sensitivity and 90% specificity. Area under the curve = 0.776; P = 0.0001.
N.R. Santos et al. / Theriogenology 71 (2009) 739–745
743
Table 2 Cox regression hazard ratios (HR) for variables affecting risk of pregnancy during the breeding season in beef cows. Variable
HR
95% CI of HR
Probability
Age group Days open Calving ease Twins Inflammation scorea
0.79 1.02 1.33 0.77 1.17
0.55–1.13 0.81–1.29 0.84–2.11 0.41–1.46 0.95–1.45
0.20 0.87 0.23 0.43 0.15
a
Fig. 6. Interactive dot diagram showing the remarkable reduction in endometritis after 50 d postpartum. Prevalence of endometritis was 82% before Day 50, and 17% after this time. Days postpartum alone predicted endometritis with 90% specificity and 71% sensitivity. (0 = endometritis absent; 1 = endometritis present.)
Because earlier pregnancy during the breeding season results in more uniform calf crops, thus conveying economic advantages [11], the influence of the same factors on the day of conception within the breeding season was examined by Cox’s proportional hazards regression. In this analysis, none of the above factors was found to influence risk (‘‘hazard’’) of conception at any time (Table 2). In dairy cows, endometrial inflammation diagnosed by cytology early in the postpartum period (before 5 wk postpartum) had little influence on subsequent reproductive performance, whereas persistence of endometritis at 7 wk postpartum or beyond was profoundly detrimental to subsequent reproduction [12]. Therefore, in the present study, cows from which samples were obtained at 50 d postpartum or later (n = 84) were analyzed separately. For this group, 15 cows (17%)
Subjective inflammation score.
were deemed to have endometritis. For cows with endometritis 13/15 (87%) became pregnant during the breeding season, compared to 64/69 (93%) for cows without evidence of endometritis (P = 0.44). The occurrence of endometritis was not influenced by age of cow, calving ease, or twinning at the previous calving. Furthermore, the median interval from the start of the breeding season to conception was 22 d for cows with endometritis after 50 d postpartum, versus 38 d for those without inflammation (P = 0.42). The proportion of macrophages also had no significant effect on proportion of cows becoming pregnant during the breeding season, or day of the breeding season on which cows became pregnant. This was true for the whole data set, and for cows examined after 50 d postpartum (considered separately). Cows that had twinned in the previous season (n = 15) were more likely to have endometritis than those giving birth to singleton calves (10/15 [67%] versus 43/120 [36%]; P = 0.02) despite the fact that all the twinning cows were more than 50 d postpartum at the time of examination. Twinning cows were more likely to be culled before the beginning of the breeding season than those that had had singleton calves (3/ 15[20%] versus 4/120 [3.3%]; P = 0.029). However, there was no difference for pregnancy proportion at the
Table 1 Multiple logistic regression for factors influencing pregnancy at the end of the breeding season in beef cows. None of the variables examined had a significant effect on pregnancy status. Variable
Coefficient a
Age group Days openb Calving ease c Twin d Inflammatione Constant a b c d e
0.616 0.092 0.301 0.900 0.148 3.356
Grouped as 2–3, 4–6, and >6 y. Stratified by quartiles. According to a 5-point scale. Dam nursing twins. Subjective inflammation score.
S.E.M.
Odds ratio
95% CI of OR
Probability
0.517 0.393 0.888 0.767 0.753 1.866
0.54 0.91 1.35 0.41 0.86
0.19–1.49 0.42–1.97 0.23–7.71 0.09–1.83 0.19–3.77
0.234 0.816 0.734 0.241 0.843 0.072
744
N.R. Santos et al. / Theriogenology 71 (2009) 739–745
end of the breeding season (9/12 [75%] versus 106/116 [91%]; P = 0.14).
4. Discussion Samples were obtained from 135/137 (99%) of cows, confirming that low-volume lavage is a practical means of obtaining diagnostic uterine samples [1,2,4]. This study also demonstrated that subjective assessment of cytological evidence of inflammation was highly repeatable between examiners, and that subjective assessment of severity of inflammation was highly correlated to the proportion of PMN (objectively determined). The nature of cytological samples was similar to that described for dairy cows [4]. In a study of >500 dairy cows in six herds, Gilbert et al. [12] reported that the incidence of cytologically diagnosed inflammation was 93% at 3 wk postpartum, 67% at 5 wk, and 51% at 7 wk. In contrast, in the present study (conducted in a single beef herd), 88% of cows with endometrial inflammation before 2 wk postpartum, 77% between 2 and 7 wk, and 17% after 7 wk. Although the proportion of affected beef cows after 50 d postpartum appeared to be less than the proportion of dairy cows, it is interesting that the proportion with demonstrable inflammation was equivalent or even higher in the preceding weeks. Gilbert et al. [13] reported that dairy cows tended to have less cytologically diagnosed endometritis after ovulation than before. In that regard, cows that ovulated earlier tended to resolve postpartum uterine inflammation more rapidly than those cows in which postpartum ovulation was delayed. In general, beef cows ovulate later postpartum than dairy cows [14,15]. This may provide an explanation for the relatively high proportion of beef cows with demonstrable endometritis before 50 d postpartum. Although the prevalence of subclinical endometritis dropped dramatically after 50 d postpartum, 17% of cows were still affected. Studies of dairy cows have shown an increase in days open and a diminished likelihood of pregnancy before 300 d postpartum for cows with cytologically diagnosed endometritis after 40 or 50 d postpartum [1,2,4,12]. In the present study, no detrimental effect of subclinical endometritis after 50 d postpartum in beef cows was found in terms of culling before the onset of the breeding season, day of season on which conception occurred, or proportion of cows pregnant by the end of the breeding season. Therefore, we inferred that beef cows efficiently resolved uterine inflammation and probably were free of it at the
initiation of the breeding season (26 d after sampling in this case). Twinning at the previous calving increased the risk of subclinical endometritis. Cows that had had twins were more likely to be culled before the onset of the breeding season, but this was not due to endometritis. Furthermore, they became pregnant in the course of the breeding season similar to nontwinning cows. It is remarkable that only one of three cows with urovagina had evidence of endometritis and the only cow with pneumovagina also had normal endometrial cytological results and that all of these cows became pregnant. It is not known what caused the clinical abnormalities, nor whether they had resolved spontaneously before the breeding season. Urovagina has been reported to cause marked endometritis [16], and it is noteworthy that cows in this study with clinically evident urovagina had no evidence of endometritis and conceived promptly during the breeding season. Pregnancy resulted from the initial insemination in 37 of 128 cows bred (29%). Although this was below expectations, it was within the range reported in the literature [17]. Older cows had higher first service conception than cows bred to calve for the second time, despite the fact that younger cows (2 and 3 y old cows) were deliberately bred to calve earlier in the preceding season; therefore, younger cows had a longer postpartum interval at the onset of the breeding season, but also had lower first-service pregnancy rates, consistent with the physiological stress of suckling a calf while still growing [18]. In conclusion, beef cows in this herd had a high prevalence of uterine inflammation before 50 d postpartum, but that appeared to be rapidly resolved thereafter (perhaps in association with postpartum ovulation). Subjective assessment of endometritis was highly correlated with objective cell counts, and in particular with the proportion of polymorphonuclear cells in cytological preparations. Cows with subclinical endometritis after 50 d postpartum had no reduction in reproductive success, including proportion culled before the beginning of the breeding season, proportion pregnant by the end of the season, or day of conception (relative to start of breeding). That this was in stark contrast to findings in dairy cows, the relevance of subclinical endometritis in beef cows appeared to be much less than for dairy cows. Because subclinical endometritis was not predictive of subsequent reproductive performance in beef cows, the use of endometrial cytology cannot be justified as a management tool in beef cows.
N.R. Santos et al. / Theriogenology 71 (2009) 739–745
Acknowledgments The authors are grateful to Ms. Heather Roman for competent technical assistance. This research was supported, in part, by the Cornell University Agricultural Experiment Station federal formula funds, Project No. NYCV-480866 received from Cooperative State Research, Education and Extension Service, U.S. Department of Agriculture. Any opinions, findings, conclusions, or recommendations expressed in this publication are those of the author(s) and do not necessarily reflect the view of the U.S. Department of Agriculture. References [1] Gilbert RO, Shin ST, Guard CL, Erb HN. Incidence of endometritis and effects on reproductive performance of dairy cows. Theriogenology 1998;49:251 [abstract]. [2] Hammon DS, Holyoak GR, Jenson J, Bingham HR. Effects of endometritis at the beginning of the breeding period on reproductive performance in dairy cows. Am Assoc Bov Pract 2001; 34:142–3. [3] Kasimanickam R, Duffield TF, Foster RA, Gartley CJ, Leslie KE, Walton JS, Johnson WH. Endometrial cytology and ultrasonography for the detection of subclinical endometritis in postpartum dairy cows. Theriogenology 2004;62:9–23. [4] Gilbert RO, Shin ST, Guard CL, Erb HN, Frajblat M. Prevalence of endometritis and its effects on reproductive performance of dairy cows. Theriogenology 2005;64:1879–88. [5] Sheldon IM, Lewis GS, LeBlanc S, Gilbert RO. Defining postpartum uterine disease in cattle. Theriogenology 2006;65:1516–30. [6] Hammon DS, Evjen IM, Dhiman TR, Goff JP. Negative energy balance during the periparturient period is associated with uterine health disorders and fever in Holstein cows. J Dairy Sci 2004;87:279 [abstract].
745
[7] LeBlanc SJ, Duffield TF, Leslie KE, Bateman KG, Keefe GP, Walton JS, Johnson WH. Defining and diagnosing postpartum clinical endometritis and its impact on reproductive performance in dairy cows. J Dairy Sci 2002;85:2223–36. [8] Hill J, Gilbert R. Reduced quality of bovine embryos cultured in media conditioned by exposure to an inflamed endometrium. Aust Vet J 2008;86:312–6. [9] Amundson JL, Mader TL, Rasby RJ, Hu QS. Environmental effects on pregnancy rate in beef cattle. J Anim Sci 2006;84:3415–20. [10] Lamb GC, Stevenson JS, Kesler DJ, Garverick HA, Brown DR, Salfen BE. Inclusion of an intravaginal progesterone insert plus GnRH and prostaglandin F2alpha for ovulation control in postpartum suckled beef cows. J Anim Sci 2001;79:2253–9. [11] Dzuik PJ, Bellows RA. Management of reproduction in beef cattle, sheep and pigs. J Anim Sci 1983;57(Suppl. 2):355 [abstract]. [12] Gilbert RO, Shin ST, Frajblat M, Guard CL, Erb HN, Roman H. The incidence of endometritis and its effect on reproductive performance of dairy cows. In: Proc 12th Int Conf Prod Dis Farm Anim, vol. 12. 2004. p. 30 [abstract]. [13] Gilbert RO, Frajblat M, Butler WR, Guard CL, Erb HN, Roman HB. The importance of early postpartum ovulation for reproduction in dairy cows. In: Proceedings, 24th world buiatrics conference; 2006 [abstract]. [14] Wheaton JE, Lamb GC. Induction of cyclicity in postpartum anestrous beef cows using progesterone, GnRH and estradiol cypionate (ECP). Anim Reprod Sci 2007;102:208–16. [15] Miller V, Ungerfeld R. Weekly bull exchange shortens postpartum anestrus in suckled beef cows. Theriogenology 2008;69:913–7. [16] Gilbert RO, Wilson DG, Levine SA, Bosu WT. Surgical management of urovagina and associated infertility in a cow. J Am Vet Med Assoc 1989;194:931–2. [17] Stevenson JS, Thompson KE, Forbes WL, Lamb GC, Grieger DM, Corah LR. Synchronizing estrus and(or) ovulation in beef cows after combinations of GnRH, norgestomet, and prostaglandin F2alpha with or without timed insemination. J Anim Sci 2000;78:1747–58. [18] Arthington JD, Kalmbacher RS. Effect of early weaning on the performance of three-year-old, first-calf beef heifers and calves reared in the subtropics. J Anim Sci 2003;81:1136–41.
Theriogenology 71 (2009) 739–745 www.theriojournal.com
Postpartum endometrial cytology in beef cows N.R. Santos a,1, G.C. Lamb b,2, D.R. Brown b, R.O. Gilbert a,* a
Department of Clinical Sciences, College of Veterinary Medicine, Cornell University, Ithaca, NY 14850, USA b North Central Research and Outreach Center, University of Minnesota, Grand Rapids, MN 55744, USA Received 12 May 2006; received in revised form 28 August 2008; accepted 5 September 2008
Abstract The objectives were to characterize postpartum endometrial cytology and to determine the prevalence of subclinical endometrial inflammation and its impact on reproduction in beef cows. Samples for endometrial cytology (low-volume uterine lavage) were obtained from 135 of 137 Angus cows (2–87 d postpartum) in northern Minnesota, 26 d before breeding started. Agreement between examiners for subjective inflammation scores was very high (kappa = 0.971); the correlation between these scores and PMN counts was high (r = 0.83; P < 0.001), validating subjective categorization. The proportion of PMN and large mononuclear cells (principally macrophages) declined with postpartum interval (P < 0.001), whereas small mononuclear cells were consistently present (and not significantly affected by postpartum interval). Pregnancy rate to fixed-time AI was 29% and overall pregnancy rate was 89%. There was no association between cell type and ultimate pregnancy status or day of conception (P > 0.10). Although inflammation later in the postpartum period apparently impaired subsequent reproduction in dairy cows, in cows >50 d postpartum at sample collection in the present study, no cytological parameter significantly predicted final pregnancy status or day of conception. Previous twinning increased the risk of subclinical endometritis (P = 0.02), but not the probability of becoming pregnant (P = 0.14). In conclusion, we inferred that beef cows had the ability to clear uterine inflammation after resumption of ovarian cyclicity. # 2009 Elsevier Inc. All rights reserved. Keywords: Endometrial cytology; Endometritis; Postpartum uterus; Fertility; Beef cows
1. Introduction Subclinical endometritis, defined by endometrial cytology, was remarkably common in dairy cows, affecting as many as 50% of all cows 7 wk postpartum
* Corresponding author at: Department of Clinical Sciences, College of Veterinary Medicine, Cornell University, Ithaca, NY 148536401, USA. Tel.: +1 607 253 3472 fax: +1 607 253 3701. E-mail address: [email protected] (R.O. Gilbert). 1 Present address: School of Veterinary Medicine, University of Pennsylvania, New Bolton Center, 382 West Street Road, Kennett Square, PA 19348, USA. 2 Present address: University of Florida, NFREC, 3925 Highway 71, Marianna, FL 32446, USA. 0093-691X/$ – see front matter # 2009 Elsevier Inc. All rights reserved. doi:10.1016/j.theriogenology.2008.09.043
in high-producing North American dairy herds [1–5]. Furthermore, dairy cows with clinical or subclinical endometritis had severely impaired subsequent reproductive performance [1,3,6,7]. The primary impact of persistent subclinical inflammation seemed to be reduced first-service pregnancy rate, which contributed to prolonged days open. Cows with persistent subclinical endometritis were also at substantially increased risk of failure to become pregnant within 300 d and being culled [4]. Inflammatory mediators had a detrimental impact on embryo development [8]. To date, no studies have examined the prevalence or consequence of subclinical endometritis in beef cows. Given that well-managed beef cow herds typically achieve high pregnancy rates (approximately 80–90%) in
740
N.R. Santos et al. / Theriogenology 71 (2009) 739–745
a limited breeding season (60–90 d) [9], we inferred that endometritis is either uncommon in beef cows, or has limited impact on subsequent reproductive performance. Nevertheless, perhaps a small number of beef cows have subclinical endometritis that impairs subsequent reproductive performance; this would present an opportunity for strategic therapeutic intervention to increase reproductive performance by curing endometritis, or for early culling if effective treatment options were not available. Therefore, our objectives were to characterize endometrial cytology, to determine the prevalence of subclinical endometritis, and to evaluate the effect of endometritis on subsequent reproduction in a single herd of beef cattle. 2. Materials and methods 2.1. Animals This study included 137 postpartum suckled Angus cows in northern Minnesota. The age of cow, date of previous calving, degree of calving difficulty, singleton or twin birth and clinical findings at the time of sample collection were recorded. Endometrial samples for diagnosis of subclinical endometritis were obtained 26 d before the beginning of the breeding season, a day selected for convenience and to provide a range of postpartum stages for examination. Cows ranged from 2 to 87 d postpartum when endometrial samples were collected. At initiation of the breeding season, ovulation was synchronized using the CO-Synch + CIDR (Controlled Internal Drug Release) protocol [10]. Briefly, cows received a 100 mg im injection of gonadotrophin releasing hormone (OvaCyst1, Phoenix Scientific, St. Joseph, MO, USA) at the time of CIDR (EAZIBREED1, Pfizer Animal Health, New York, NY, USA) insertion and a 25 mg injection of prostaglandin F2a (Lutalyse1, Pfizer Animal Health) at CIDR removal 7 d later, followed in 48 h by a second injection of GnRH and fixed-time AI. After this initial insemination, cows were inseminated by one of the three experienced AI technicians at observed estrus during a 70-d season. Estrus was observed three times daily for at least 30 min; cows were inseminated within 12 h of detection of standing estrus. Pregnancy was diagnosed at 28–38 d after AI by transrectal ultrasonography (5-MHz intrarectal linear-array transducer, Aloka 500 V, Corometrics, Wallingford, CT, USA) and confirmed at 58– 68 d of pregnancy by a single individual. The presence or absence of pregnancy and date of successful insemination were recorded.
2.2. Sample collection and analysis Samples were obtained by aseptic infusion of 20 mL of sterile Dulbecco’s phosphate buffered saline solution via a plastic infusion catheter (MAI/Genesis 21 in flex tip pipette, Butler Animal Health Supply, Dublin, OH, USA), followed by gentle uterine agitation and aspiration of approximately 5 mL of fluid. Recovered fluid was processed within 6 h after collection by brief vortexing and centrifugation onto a glass slide in a cytocentrifuge (Cytotek cytocentrifuge, 70 g, 10 min). Slides were air dried, stained with a modified Wright-Giemsa stain (Harleco1 Hemacolor1; EMD Chemicals, Gibbstown, NJ, USA), a cover slip was attached with Cytoseal 601 (Richard-Allan Scientific, Kalamazoo, MI, USA) and slides examined under bright-field microscopy at 200 and 400. Each cow was categorized as positive or negative for endometritis by two independent examiners, using subjective assessment of the cytological preparation, as described by Gilbert et al. [4]. A subjective score of 0– 3 was assigned independently by each observer, with 0 representing the absence of inflammation and 3 severe inflammation. Independently, a third examiner counted 200 cells from each slide, categorizing them as epithelial cells, polymorphonuclear cells, large mononuclear cells (presumed to be macrophages), and small mononuclear cells (presumed to be lymphocytes). The proportion of each was calculated. All examiners were unaware of demographic information regarding each cow. 2.3. Statistical analysis The degree of agreement between examiners of the cytological preparations was determined by the calculation of the kappa statistic. The relationship between subjective assessment of endometritis and proportions of counted cells was ascertained by nonparametric analysis of variance (Kruskal–Wallis test). Pearson correlations between proportions of cell types were calculated. The effect of presence or absence and degree of inflammation, and the proportions of each cell type in cytological preparations, calving difficulty, twinning at the previous calving, age and days open at the beginning of the breeding season on the probability of becoming pregnant, and the day of the breeding season on which conception occurred was examined by multiple logistic regression and Cox’s proportional hazards regression. Differences in proportions of cows pregnant at the end of the breeding season were examined by Fishers exact test. Changes in cell populations in cytological preparations with days
N.R. Santos et al. / Theriogenology 71 (2009) 739–745
741
postpartum were examined by analysis of variance and fitted to a polynomial regression curve. Optimal cutoff points for determining the relationship between subjectively diagnosed inflammation and percentage of neutrophils in cytology samples, and for exploring the relationship between days postpartum and presence of inflammation were calculated using receiver operating characteristic (ROC) curves. Specificity, sensitivity and P-value of the ROC test are reported. Statistical calculations were performed using SigmaStat for Windows, Version 3.0 (SPSS, Inc., Chicago, IL, USA), MedCalc for Windows, version 9.6.0.0 (MedCalc Software, Mariakerke, Belgium) and StataIC 10.0 (StataCorp LP, College Station, TX, USA) software packages. 3. Results 3.1. Validation of technique Samples were obtained from 135 of 137 cows (99%); in two cases difficulty in passing the pipette through the cervix or recovering lavage fluid precluded obtaining a diagnostic sample. Weighted kappa score for agreement between examiners for assignment of subjective inflammation category was 0.971. Cows ranged from 2 to 87 d postpartum at the time of examination. Subclinical endometritis (subjectively determined) was diagnosed in 88% of cows in the first quartile (2–23 d postpartum); 34% of cows in the second quartile (24– 57 d postpartum); 19% in the third quartile (58–73 d) and 19% in the fourth quartile (P < 0.0001). The correlation between subjective inflammation score and proportion of polymorphonuclear cells in the cytological preparations was 0.83 (r2 = 0.69, P < 0.001; Fig. 1). There were differences between the groups (P < 0.001), with significant differences between Group 0 (no inflammation) and all the other groups, but not among degrees of inflammation. Based on an ROC curve, a cutoff point of >5.5% polymorphonuclear cells was highly correlated with subjective assignment as inflamed; this cutoff point was 100% specific and 78% sensitive for predicting subjectively assigned inflammation score of 1 or greater (Figs. 2 and 3). The proportions of polymorphonuclear cells and large mononuclear cells were independent of each other (r < 0.01, P > 0.9). Similarly, the proportions of large and small mononuclear cells were independent of each other (r = 0.1, P = 0.9). However, the proportions of polymorphonuclear cells and small mononuclear cells were negatively correlated (r = 0.2, P = 0.037).
Fig. 1. Percentage of polymorphonuclear cells in cytology preparations (of endometrial lavage) by subjective inflammation scores (difference among inflammation scores, P < 0.001). This figure shows median PMN%, the interquartile range, the 5th and 95th percentile, and individual outliers.
Fig. 2. Receiver Operating Characteristic (ROC) curve for percentage of polymorphonuclear cells as a predictor of subjectively assigned inflammation in beef cows. Analysis of the ROC indicated that the use of a cutoff point of more than 5.5% PMNs was 100% specific and 78% sensitive for inflammation as assessed subjectively. Area under the curve = 0.951, P = 0.0001.
742
N.R. Santos et al. / Theriogenology 71 (2009) 739–745
Fig. 3. This interactive dot diagram depicts proportion of polymorphonuclear cells in samples subjectively assessed as positive (1) or negative (0) for inflammation. Note that none of the samples judged not to be inflamed had PMN proportions greater than 5.5%. A few samples subjectively judged to be inflamed, however, had PMN proportions lower than 5.5%.
3.2. Prevalence and impact of subclinical endometritis The proportion of PMN cells declined with postpartum interval (r = 0.57, P < 0.0001). This decline most closely matched a quadratic curve (PMN% = 39 1.08 DPP + 0.008 DPP2, where DPP = days postpartum; r2 = 0.35, P < 0.0001; Fig. 4). The proportion of cows with cytologically diagnosed endometritis after 50 d postpartum was particularly low (Figs. 5 and 6). The proportion of macrophages also declined with increasing postpartum interval (P = 0.0001), but this was not true for lymphocytes (P = 0.49). Seven cows were sold after sampling but before the onset of the breeding season. None of these cows had cytological evidence of endometritis. At the time of sampling three cows were found to have severe urovagina; only one of these had evidence of endometritis. One cow had marked pneumovagina, but also lacked evidence of endometritis. All of these cows became pregnant, three on the first day of the breeding season, and the fourth 22 d after the onset of the breeding season. Pregnancy resulted from the initial insemination in 37 of 128 cows bred (29%). No cytological feature affected pregnancy to first service (P > 0.4). Pregnancy to first insemination was 28% for cows aged 2 or 3 y, 34% for those 4–6 y, and 21% for cows 7 y (P = 0.21). Days to onset of the breeding season, stratified by quartile, were examined as a potential factor affecting pregnancy to first service, but was not significant (P = 0.68).
Fig. 4. The proportion of PMN cells declined with increasing days postpartum. (PMN% = 39 1.08 DPP + 0.08 DPP2, where DPP = days postpartum; r2 = 0.35; P < 0.0001).
None of the measured factors (age, days open at onset of breeding season, calving ease, previous twinning, subjective diagnosis of endometritis, or proportion of any cell type) significantly influenced the likelihood of pregnancy at the end of the breeding season (Table 1.)
Fig. 5. Receiver Operating Characteristic (ROC) curve for the presence of inflammation by days postpartum. Using a cutoff point of 50 d postpartum predicted endometritis with 71% sensitivity and 90% specificity. Area under the curve = 0.776; P = 0.0001.
N.R. Santos et al. / Theriogenology 71 (2009) 739–745
743
Table 2 Cox regression hazard ratios (HR) for variables affecting risk of pregnancy during the breeding season in beef cows. Variable
HR
95% CI of HR
Probability
Age group Days open Calving ease Twins Inflammation scorea
0.79 1.02 1.33 0.77 1.17
0.55–1.13 0.81–1.29 0.84–2.11 0.41–1.46 0.95–1.45
0.20 0.87 0.23 0.43 0.15
a
Fig. 6. Interactive dot diagram showing the remarkable reduction in endometritis after 50 d postpartum. Prevalence of endometritis was 82% before Day 50, and 17% after this time. Days postpartum alone predicted endometritis with 90% specificity and 71% sensitivity. (0 = endometritis absent; 1 = endometritis present.)
Because earlier pregnancy during the breeding season results in more uniform calf crops, thus conveying economic advantages [11], the influence of the same factors on the day of conception within the breeding season was examined by Cox’s proportional hazards regression. In this analysis, none of the above factors was found to influence risk (‘‘hazard’’) of conception at any time (Table 2). In dairy cows, endometrial inflammation diagnosed by cytology early in the postpartum period (before 5 wk postpartum) had little influence on subsequent reproductive performance, whereas persistence of endometritis at 7 wk postpartum or beyond was profoundly detrimental to subsequent reproduction [12]. Therefore, in the present study, cows from which samples were obtained at 50 d postpartum or later (n = 84) were analyzed separately. For this group, 15 cows (17%)
Subjective inflammation score.
were deemed to have endometritis. For cows with endometritis 13/15 (87%) became pregnant during the breeding season, compared to 64/69 (93%) for cows without evidence of endometritis (P = 0.44). The occurrence of endometritis was not influenced by age of cow, calving ease, or twinning at the previous calving. Furthermore, the median interval from the start of the breeding season to conception was 22 d for cows with endometritis after 50 d postpartum, versus 38 d for those without inflammation (P = 0.42). The proportion of macrophages also had no significant effect on proportion of cows becoming pregnant during the breeding season, or day of the breeding season on which cows became pregnant. This was true for the whole data set, and for cows examined after 50 d postpartum (considered separately). Cows that had twinned in the previous season (n = 15) were more likely to have endometritis than those giving birth to singleton calves (10/15 [67%] versus 43/120 [36%]; P = 0.02) despite the fact that all the twinning cows were more than 50 d postpartum at the time of examination. Twinning cows were more likely to be culled before the beginning of the breeding season than those that had had singleton calves (3/ 15[20%] versus 4/120 [3.3%]; P = 0.029). However, there was no difference for pregnancy proportion at the
Table 1 Multiple logistic regression for factors influencing pregnancy at the end of the breeding season in beef cows. None of the variables examined had a significant effect on pregnancy status. Variable
Coefficient a
Age group Days openb Calving ease c Twin d Inflammatione Constant a b c d e
0.616 0.092 0.301 0.900 0.148 3.356
Grouped as 2–3, 4–6, and >6 y. Stratified by quartiles. According to a 5-point scale. Dam nursing twins. Subjective inflammation score.
S.E.M.
Odds ratio
95% CI of OR
Probability
0.517 0.393 0.888 0.767 0.753 1.866
0.54 0.91 1.35 0.41 0.86
0.19–1.49 0.42–1.97 0.23–7.71 0.09–1.83 0.19–3.77
0.234 0.816 0.734 0.241 0.843 0.072
744
N.R. Santos et al. / Theriogenology 71 (2009) 739–745
end of the breeding season (9/12 [75%] versus 106/116 [91%]; P = 0.14).
4. Discussion Samples were obtained from 135/137 (99%) of cows, confirming that low-volume lavage is a practical means of obtaining diagnostic uterine samples [1,2,4]. This study also demonstrated that subjective assessment of cytological evidence of inflammation was highly repeatable between examiners, and that subjective assessment of severity of inflammation was highly correlated to the proportion of PMN (objectively determined). The nature of cytological samples was similar to that described for dairy cows [4]. In a study of >500 dairy cows in six herds, Gilbert et al. [12] reported that the incidence of cytologically diagnosed inflammation was 93% at 3 wk postpartum, 67% at 5 wk, and 51% at 7 wk. In contrast, in the present study (conducted in a single beef herd), 88% of cows with endometrial inflammation before 2 wk postpartum, 77% between 2 and 7 wk, and 17% after 7 wk. Although the proportion of affected beef cows after 50 d postpartum appeared to be less than the proportion of dairy cows, it is interesting that the proportion with demonstrable inflammation was equivalent or even higher in the preceding weeks. Gilbert et al. [13] reported that dairy cows tended to have less cytologically diagnosed endometritis after ovulation than before. In that regard, cows that ovulated earlier tended to resolve postpartum uterine inflammation more rapidly than those cows in which postpartum ovulation was delayed. In general, beef cows ovulate later postpartum than dairy cows [14,15]. This may provide an explanation for the relatively high proportion of beef cows with demonstrable endometritis before 50 d postpartum. Although the prevalence of subclinical endometritis dropped dramatically after 50 d postpartum, 17% of cows were still affected. Studies of dairy cows have shown an increase in days open and a diminished likelihood of pregnancy before 300 d postpartum for cows with cytologically diagnosed endometritis after 40 or 50 d postpartum [1,2,4,12]. In the present study, no detrimental effect of subclinical endometritis after 50 d postpartum in beef cows was found in terms of culling before the onset of the breeding season, day of season on which conception occurred, or proportion of cows pregnant by the end of the breeding season. Therefore, we inferred that beef cows efficiently resolved uterine inflammation and probably were free of it at the
initiation of the breeding season (26 d after sampling in this case). Twinning at the previous calving increased the risk of subclinical endometritis. Cows that had had twins were more likely to be culled before the onset of the breeding season, but this was not due to endometritis. Furthermore, they became pregnant in the course of the breeding season similar to nontwinning cows. It is remarkable that only one of three cows with urovagina had evidence of endometritis and the only cow with pneumovagina also had normal endometrial cytological results and that all of these cows became pregnant. It is not known what caused the clinical abnormalities, nor whether they had resolved spontaneously before the breeding season. Urovagina has been reported to cause marked endometritis [16], and it is noteworthy that cows in this study with clinically evident urovagina had no evidence of endometritis and conceived promptly during the breeding season. Pregnancy resulted from the initial insemination in 37 of 128 cows bred (29%). Although this was below expectations, it was within the range reported in the literature [17]. Older cows had higher first service conception than cows bred to calve for the second time, despite the fact that younger cows (2 and 3 y old cows) were deliberately bred to calve earlier in the preceding season; therefore, younger cows had a longer postpartum interval at the onset of the breeding season, but also had lower first-service pregnancy rates, consistent with the physiological stress of suckling a calf while still growing [18]. In conclusion, beef cows in this herd had a high prevalence of uterine inflammation before 50 d postpartum, but that appeared to be rapidly resolved thereafter (perhaps in association with postpartum ovulation). Subjective assessment of endometritis was highly correlated with objective cell counts, and in particular with the proportion of polymorphonuclear cells in cytological preparations. Cows with subclinical endometritis after 50 d postpartum had no reduction in reproductive success, including proportion culled before the beginning of the breeding season, proportion pregnant by the end of the season, or day of conception (relative to start of breeding). That this was in stark contrast to findings in dairy cows, the relevance of subclinical endometritis in beef cows appeared to be much less than for dairy cows. Because subclinical endometritis was not predictive of subsequent reproductive performance in beef cows, the use of endometrial cytology cannot be justified as a management tool in beef cows.
N.R. Santos et al. / Theriogenology 71 (2009) 739–745
Acknowledgments The authors are grateful to Ms. Heather Roman for competent technical assistance. This research was supported, in part, by the Cornell University Agricultural Experiment Station federal formula funds, Project No. NYCV-480866 received from Cooperative State Research, Education and Extension Service, U.S. Department of Agriculture. Any opinions, findings, conclusions, or recommendations expressed in this publication are those of the author(s) and do not necessarily reflect the view of the U.S. Department of Agriculture. References [1] Gilbert RO, Shin ST, Guard CL, Erb HN. Incidence of endometritis and effects on reproductive performance of dairy cows. Theriogenology 1998;49:251 [abstract]. [2] Hammon DS, Holyoak GR, Jenson J, Bingham HR. Effects of endometritis at the beginning of the breeding period on reproductive performance in dairy cows. Am Assoc Bov Pract 2001; 34:142–3. [3] Kasimanickam R, Duffield TF, Foster RA, Gartley CJ, Leslie KE, Walton JS, Johnson WH. Endometrial cytology and ultrasonography for the detection of subclinical endometritis in postpartum dairy cows. Theriogenology 2004;62:9–23. [4] Gilbert RO, Shin ST, Guard CL, Erb HN, Frajblat M. Prevalence of endometritis and its effects on reproductive performance of dairy cows. Theriogenology 2005;64:1879–88. [5] Sheldon IM, Lewis GS, LeBlanc S, Gilbert RO. Defining postpartum uterine disease in cattle. Theriogenology 2006;65:1516–30. [6] Hammon DS, Evjen IM, Dhiman TR, Goff JP. Negative energy balance during the periparturient period is associated with uterine health disorders and fever in Holstein cows. J Dairy Sci 2004;87:279 [abstract].
745
[7] LeBlanc SJ, Duffield TF, Leslie KE, Bateman KG, Keefe GP, Walton JS, Johnson WH. Defining and diagnosing postpartum clinical endometritis and its impact on reproductive performance in dairy cows. J Dairy Sci 2002;85:2223–36. [8] Hill J, Gilbert R. Reduced quality of bovine embryos cultured in media conditioned by exposure to an inflamed endometrium. Aust Vet J 2008;86:312–6. [9] Amundson JL, Mader TL, Rasby RJ, Hu QS. Environmental effects on pregnancy rate in beef cattle. J Anim Sci 2006;84:3415–20. [10] Lamb GC, Stevenson JS, Kesler DJ, Garverick HA, Brown DR, Salfen BE. Inclusion of an intravaginal progesterone insert plus GnRH and prostaglandin F2alpha for ovulation control in postpartum suckled beef cows. J Anim Sci 2001;79:2253–9. [11] Dzuik PJ, Bellows RA. Management of reproduction in beef cattle, sheep and pigs. J Anim Sci 1983;57(Suppl. 2):355 [abstract]. [12] Gilbert RO, Shin ST, Frajblat M, Guard CL, Erb HN, Roman H. The incidence of endometritis and its effect on reproductive performance of dairy cows. In: Proc 12th Int Conf Prod Dis Farm Anim, vol. 12. 2004. p. 30 [abstract]. [13] Gilbert RO, Frajblat M, Butler WR, Guard CL, Erb HN, Roman HB. The importance of early postpartum ovulation for reproduction in dairy cows. In: Proceedings, 24th world buiatrics conference; 2006 [abstract]. [14] Wheaton JE, Lamb GC. Induction of cyclicity in postpartum anestrous beef cows using progesterone, GnRH and estradiol cypionate (ECP). Anim Reprod Sci 2007;102:208–16. [15] Miller V, Ungerfeld R. Weekly bull exchange shortens postpartum anestrus in suckled beef cows. Theriogenology 2008;69:913–7. [16] Gilbert RO, Wilson DG, Levine SA, Bosu WT. Surgical management of urovagina and associated infertility in a cow. J Am Vet Med Assoc 1989;194:931–2. [17] Stevenson JS, Thompson KE, Forbes WL, Lamb GC, Grieger DM, Corah LR. Synchronizing estrus and(or) ovulation in beef cows after combinations of GnRH, norgestomet, and prostaglandin F2alpha with or without timed insemination. J Anim Sci 2000;78:1747–58. [18] Arthington JD, Kalmbacher RS. Effect of early weaning on the performance of three-year-old, first-calf beef heifers and calves reared in the subtropics. J Anim Sci 2003;81:1136–41.