Safety of a progesterone-releasing intravaginal device as assessed from vaginal mucosal integrity and indicators of systemic inflammation in postpartum dairy cows

Abstract

A clinical trial was conducted to investigate the animal safety of a progesterone-releasing intravaginal device (PRID). Anestrus cows at a mean of 63 ± 3.5 d in milk were randomly assigned to an ovulation-synchronization protocol that included placement of either a PRID or a placebo intravaginal device (PID) or no such treatment. At enrolment and at device removal 7 d later, blood samples were collected. The outcomes of interest included the vaginal reaction to the device, the vaginal mucosal integrity, and the results of bacterial culture of swabs of the vaginal mucosa. In addition, the leukocyte and haptoglobin responses were measured. Although only 5% of the PRID-treated animals compared with 19% of the PID-treated animals had a copious purulent vaginal discharge at the time of device removal, there was no significant difference in the proportions; furthermore, there was no evidence of vaginal mucosal damage associated with either device. The total blood leukocyte count was significantly lower in both the PRID-treated cows and the PID-treated cows after device removal compared with the start of treatment (P < 0.05) and compared with no treatment (P < 0.001); there was no difference in leukocyte response between the 2 device-treated groups. The decrease in leukocyte count was attributed to a significant reduction in the numbers of circulating neutrophils and lymphocytes, a pattern consistent with the luteal phase of the bovine estrus cycle. There was no significant difference in the circulating haptoglobin concentration between the 3 groups of cows. Culture revealed commensal bacterial growth in the vagina of all the cows.

Introduction

Reproductive hormonal therapy is increasingly common on modern dairy farms (1). Adoption of ovulation-synchronization programs is predominantly driven by their proven efficacy and ease of implementation. However, the probability of pregnancy after 1st insemination in a fixed-time insemination program is 5% to 20% lower in cows without a predictable ovulatory cycle by the end of the voluntary waiting period relative to cyclic cows (2,3). Inclusion of a progesterone-releasing intravaginal device (PRID) has significantly improved synchronization of estrus and the probability of pregnancy after 1st insemination in anovulatory cows (3–5). Despite the efficacy of progesterone supplementation in anestrus anovulatory cows, there are few published reports that address animal health and safety with the use of this technology.

The PRID (Vétoquinol, Lavaltrie, Quebec) is a Silastic coil approximately 8 cm long, with an outside diameter of 4.5 cm, which is impregnated with 1.55 g of progesterone. At the time of insertion, the device is tightly coiled around the outside of the insertion tool, which reduces the outside diameter to 2.5 cm. The PRID is retained on the insertion tool by being hooked onto a caudal-facing hook at the rostral end, as well as by a round peg and release mechanism at the caudal end. There are no sharp edges on the PRID or the insertion tool. Device retention relies on resistance associated with the broad, flat surface of the PRID against the vaginal mucosa. This large area of contact ensures appropriate release of the natural progesterone over the 7-d exposure period. A 30-cm-long cord attached to the caudal end of the PRID is left protruding from the vulva. The device is removed by gently pulling the exposed cord.

Subjective assessments of the associated vaginal reaction have been reported (6), but descriptions of hematologic responses have not been reported. The immune response is monitored through differential leukocyte counts and the acute-phase protein response. Differential leukocyte counts include the relative contribution of neutrophils, monocytes, lymphocytes, eosinophils, and basophils to the circulating pool (7). Haptoglobin, an acute-phase protein produced by the liver when stimulated by cytokines, is associated with infection and has been used as an indicator of mastitis, metritis, and retained placenta but not of chronic endometritis (8).

The objective of this investigation was to assess the impact of a 7-d PRID treatment protocol on vaginal mucosal integrity as evaluated by vaginoscopy, on the local immune response as measured by a vaginitis score, and on the systemic immune response as measured by the response in both the circulating leukocyte count and the serum haptoglobin concentration.

Materials and methods

Animals

This study was conducted among Holstein cows at a large free-stall dairy farm close to Guelph, Ontario. Animals were selected for enrolment on the basis of failure to display estrus by 63 ± 3.5 d in milk, as determined by the absence of an increase in walking activity, measured by a pedometer (AfiAct System; Germania Dairy Automation, Waunakee, Wisconsin, USA). All animals were cared for in accordance with the Canadian Council on Animal Care guidelines (9), and the experimental protocol (04R014) was approved by the Animal Care Committee of the University of Guelph.

Experimental protocol

All eligible cows were enrolled in an ovulation-synchronization protocol (10) at weekly herd visits conducted by the lead investigator. On study day 0, all cows were given an intramuscular (IM) injection of 100 μg of gonadorelin acetate (Fertiline; Vétoquinol) and then received, by random assignment, a placebo intravaginal device (PID; group A), a PRID (group B), or no intravaginal device (negative-control group C). The PID was constructed of the same Silastic coil as the PRID, except that it was removed from the manufacturing process before progesterone impregnation. Before device insertion, the vulva was washed with warm water and dried with a clean towel. A small amount of 1% (w/w) chlorhexidine ointment (Hibitane Veterinary Ointment; Wyeth Animal Health, Guelph, Ontario) was applied to the device before insertion. On study day 7, the intravaginal device was removed from groups A and B, and all cows received an IM injection of 500 μg of cloprostenol sodium (Estrumate; Schering-Plough Animal Health, Pointe Claire, Quebec), a synthetic prostaglandin. Fixed-time insemination was performed on study day 10, approximately 16 h after a 2nd injection of gonadorelin acetate, on study day 9.

Vaginal health assessment

Before device removal (on study day 7), the vulva was washed with warm water, a sterile bacterial culture swab was inserted along the dorsal vaginal wall to approximately 10 cm rostral to the vulva, and a swab of the vaginal mucosa was taken. The intravaginal device was then removed. A vaginitis score was assigned according to the amount and type of debris associated with the device (0 — no debris; 1 — small flecks of purulent debris on the device; 2 — copious amounts of purulent debris on the device and vulva). Subsequently, a vaginoscopic examination was conducted and a score assigned to the vaginal mucosal integrity (0 — no visible lesions; 1 — superficial lesions; 2 — erosions of the vaginal mucosa) in all 3 groups of cows.

The swabs were transported to the laboratory in bacterial culturettes (BBL Culture Swab Collection and Transportation System; Becton, Dickinson and Company, Sparks, Maryland, USA) and plated on MacConkey’s agar and blood agar with 1% esculin. Coliform bacteria were diagnosed on the basis of growth on MacConkey’s agar. Colonies growing on blood agar were distinguished as staphylococcal or streptococcal according to the results of the catalase test. The total number of colonies on the plates was scored on a 4-point scale (+ — 1 to 5; + +— 6 to 10; + + + — 11 to 50; + + + + — > 50 colonies).

Blood sampling and analysis

On study days 0 and 7, a blood sample was collected from each cow by coccygeal venipuncture and permitted to clot for a minimum of 60 min in a BD Vacutainer (Becton Dickinson, Franklin Lakes, New Jersey, USA). Another blood sample was collected into a BD Vacutainer containing ethylene diamine tetraacetic acid. Both samples were transported to the laboratory within 2 h of collection. The clotted sample was centrifuged at 375 × g for 10 min, and the serum was separated and stored at −20°C pending laboratory analysis. The haptoglobin concentration in the serum was determined with the Hitachi 911 automated analyzer (Roche Diagnostics, Indianapolis, Indiana, USA), which measures the hemoglobin-binding capacity using a methemoglobin reagent made by the University of Guelph Animal Health Laboratory (AHL) according to the method described by Skinner and Roberts (11). Haptoglobin analysis was done at the end of the study period. The whole-blood sample underwent a complete leukocyte count and an automated differential count with the ADVIA 120 (Bayer Diagnostics, Tarrytown, New York, USA), which were compared with the AHL reference limits (12).

Exclusions

Data were excluded from analysis for each cow in groups A and B whose device was absent on study day 7 and each cow in any of the 3 groups that displayed estrus (increased pedometer activity) between study days 0 and 7.

Data management and statistical analysis

Hematologic values and bacterial culture data were entered into a database weekly. Within a treatment group, a 2-tailed paired Student’s t test was applied to assess differences in the analyte at enrolment versus 7 d later, at device removal. Between treatment groups, the difference between enrolment and 7 d later was assessed by multivariable linear regression in which the analyte concentration at enrolment was used as a covariate to predict the final concentration. All models were repeated with data excluded that were deemed outliers: large residual values, unusual observations, and those exerting undue influence. If the estimate was significantly altered by the exclusions, the range of effect is presented. Modelling assumptions of equal variance and normality of residuals were evaluated graphically. In all models, the assumption of normality of residuals was violated. A series of power transformations, performed with Intercooled Stata statistical software (release 9.1, 2005; Stata Corporation, College Station, Texas, USA) did not solve the lack of normality and did not alter the inferences about the significance of variables in the model. Therefore, for ease of interpretation, the raw data were retained in the final model.

Categorical data were analyzed by means of the χ² test, with application of Fisher’s exact test for estimation of probability values to account for the small sample size. Data for animals outside the published reference range are discussed separately.

Results

Of the 65 animals enrolled in the study, 26 received a PID, 24 a PRID, and 15 no intravaginal device. During the 7-d insertion period, 7, 2, and 2 animals in groups A, B, and C, respectively, displayed increased pedometer activity. In addition, 3 PIDs and 2 PRIDs were lost before the prescribed time for removal; the rates of device retention were therefore 84% for the PID and 91% for the PRID. Thus, 16 PID-treated cows, 20 PRID-treated cows, and 13 negative-control cows remained for analysis.

Table I reports the vaginitis scores. Copious purulent debris was observed on the device at removal in 3 (19%) of the 16 PID recipients versus 1 (5%) of the 20 PRID recipients. The difference was not statistically significant (P = 0.26). Vaginoscopic examination on day 7 showed no evidence of vaginal mucosal damage in any animal in the 3 groups. There was no cloudy mucus at fixed-time insemination, 3 d after device removal.

Table I.

Vaginitis score in Holstein cows at the time of removal of a placebo intravaginal device (PID) or a progesterone-releasing intravaginal device (PRID) that had been in place for 7 d

	Vaginitis score;b no. (and %) of cows
Treatmenta	0	1	2
PID	9/16 (56)	4/16 (25)	3/16 (19)
PRID	13/19 (68)	5/19 (26)	1/19 (5)

^aBoth groups, as well as the negative-control group, were given an intramuscular injection of 100 μg of gonadorelin acetate on study day 0, before device insertion in the case of the PID and PRID groups.

^bAssigned according to the amount and type of debris associated with the device: 0 — no debris; 1 — small flecks of purulent debris on the device; 2 — copious amounts of purulent debris on the device and vulva.

There were 3 abnormal situations. One animal received a device despite a diagnosis of pyometra. In this animal, the leukocyte count remained within normal limits, the neutrophil count remained below the lower limit of the reference interval, and the serum haptoglobin concentrations were within the reference interval. A uterine discharge was evident at device removal. A 2nd animal displayed estrus during the 7-d treatment period. According to protocol, the device was removed and the animal inseminated. However, the owner reported a foul odour associated with the device at removal. Yet vaginoscopic examination at the scheduled visit did not identify an abnormality. In the 3rd case, the device rotated 180° within the vagina, such that the attached string was no longer visible externally. The device was identified via palpation per rectum and was removed uneventfully. There was a copious amount of flocculent material at the time of device removal.

Figure 1A displays the mean total leukocyte count at enrolment and 7 d later for the 3 groups. Within treatments, the count decreased over the 7 d in the groups that received a device (P < 0.05). The animals treated with a device had a greater decrease in the count (P < 0.001) than the animals not treated with a device (Table II). There was no difference in the leukocyte response between groups A and B. The total leukocyte count for 1 PID and 1 PRID recipient was above the reference range at the start of treatment and returned to the reference range by the end of treatment. Two animals in the negative-control group had elevated total leukocyte counts throughout the treatment period. Exclusion of these abnormal observations did not alter the interpretation of the data. Despite the significant differences, all values were within reference intervals suggested by the AHL.

Figure 1.

Comparison of mean (and standard error) of total leukocyte (A), neutrophil (B), monocyte (C), and lymphocyte (D) counts in lactating Holstein dairy cattle at study enrolment (white bars), at 63 ± 3.5 d in milk, and 7 d (grey bars) after being given an intramuscular injection of 100 μg of gonadorelin acetate and a placebo intravaginal device (PID), a progesterone-releasing intravaginal device (PRID), or no intravaginal device (negative control). Asterisks indicate a significant difference (P < 0.05) over time, daggers a significant difference (P < 0.05) from the control value.

Table II.

Multivariable logistic regression of the impact of treatment with an intravaginal device with or without progesterone relative to no device treatment (in the negative-control group) on the total leukocyte and differential counts and the serum haptoglobin concentration, controlling for the analyte concentration at enrolment

Variable	Treatment	Coefficient	P	95% confidence interval
Total leukocyte count	PID	−1.98	< 0.001	−3.0 to −0.95
	PRID	−1.92	< 0.001	−2.9 to −0.92
	Neither	Referent
Neutrophil count	PID	−0.82	0.02	−1.5 to −0.14
	PRID	−1.11	0.002	−1.8 to −0.45
	Neither	Referent
Monocyte count	PID	−0.01	0.8	−0.2 to 0.14
	PRID	−0.14	0.08	−0.3 to 0.02
	Neither	Referent
Lymphocyte count	PID	−1.40	0.001	−2.2 to −0.64
	PRID	−0.14	0.02	−1.7 to −0.18
	Neither	Referent
Haptoglobin concentration	PID	−0.33	0.23	−0.9 to 0.21
	PRID	−0.27	0.31	−0.8 to 0.27
	Neither	Referent

Figures 1B to 1D display the mean segmented neutrophil, monocyte, and lymphocyte counts at enrolment and 7 d later. Within treatments, the neutrophil count decreased over the 7 d in the groups that received a device (P < 0.05), the monocyte count did not change significantly, and the lymphocyte count decreased in the PID-treated animals (P = 0.008), did not change in the PRID-treated animals, and increased in the negative-control group (P = 0.04). The presence of a device reduced the neutrophil and lymphocyte counts (P ≤ 0.02) relative to those of the negative-control group (Table II). One animal in each treatment group had an elevated segmented neutrophil count at the beginning of treatment, but in none was the count elevated at the end of treatment. Exclusion of these animals did not significantly alter the interpretation. The presence of progesterone in the intravaginal device did not alter the final neutrophil or monocyte count (Table II). The lymphocyte count rose above the reference range during treatment in 1 PID and 2 PRID recipients.

Within treatments, there were no significant differences in the serum concentration of haptoglobin from enrolment to device removal (Figure 2). Between treatments, there was no significant difference in the serum haptoglobin concentration at device removal (Table II).

Figure 2.

Comparison of the haptoglobin response in the 3 groups of cows. There were no significant differences within or between treatments.

Table III presents, by isolate, the results of bacterial culture of the swabs of the dorsal vaginal wall approximately 10 cm rostral to the vulva in the 14, 17, and 11 cows in groups A, B, and C, respectively, for which results were available for analysis. A single isolate was cultured from 9, 9, and 8 animals, respectively, and mixed isolates were cultured from the remaining 5, 8, and 3 animals. The median number of colonies per plate, independent of organism cultured, was > 50 for all 3 groups.

Table III.

Bacterial culture results for swabs of the dorsal vaginal wall 10 cm rostral to the vulva on study day 7, before device removal

	Treatment; no. of cows with isolate (and estimate of growtha)
Bacterial isolate	PID n ( = 14)	PRID ( n = 17)	Neither n ( = 11)
Arcanobacterium pyogenes	1 (+ + + +)	1 (+ + + + )	0
Gram-positive rods	2 (+ + + )	1 (+)	2 (+ + +)
Streptococci	2 (+ + + +)	0	3 (+ + +)
Coliforms	2 (+ + + )	2 (+ + + )	3 (+ + + +)
Proteus sp.	2 (+++)	5 (++++ )	0
Mixedb	5 (+ + + + )	8 (+ + + + )	3 (+ + + +)

^aScored on a 4-point scale according to the median number of colonies: + — 1 to 5; + + — 6 to 10; + + + — 11 to 50; + + + + — > 50.

^bCombination of Staphylococcus sp. in addition to the organisms listed above.

Discussion

Retention rates for the PRID have been reported to range from 88% in housed dairy cows to 96% in dairy cows at pasture (13); the rate in our study was 91%. A 97% rate has been reported for a T-shaped controlled-release intravaginal device (14). The broad surface of the PRID is thought to increase the likelihood of device retention. In this study, chlorhexidine ointment was used as an antibacterial lubricant. Use of creams has been associated with increased likelihood of loss of the device (15).

Vaginal reaction to device insertion has been consistent in recent studies (16,17). Vaginoscopic examination after device removal was unique to this investigation. All cows fitted with an intravaginal device had some debris visible caudal to the cervical os. Cows that had significant amounts of cloudy mucus associated with the device at removal also had an accumulation of mucus in the vagina. The lack of power in the current study precludes statistical analysis of these data. However, in a larger study, cows treated with a PRID were 60% less likely to have a vaginitis score of 2 than PID-treated cows (7.7% versus 17.2%; odds ratio 0.4; P < 0.001) (17). The use of a placebo device has rarely been incorporated into large-scale reproductive trials to permit separation of the effect of the device and the effect of progesterone. Vaginitis assessments have been published for both the PRID and the progesterone-insert Eazi-Breed CIDR (controlled internal drug release) Cattle Device (Pfizer Animal Health, Pfizer Canada, Kirkland, Quebec) in postinsemination applications (6,14). A 14-d PRID treatment induced localized vaginitis in approximately 50% of treated animals without affecting the likelihood of pregnancy (6). Similarly, 65% of animals had a mild to moderate vaginal reaction to a 7-d exposure to the CIDR, 27% of the devices being associated with yellow mucus at removal (14).

The vaginitis scores in our study are the result of the combined use of 1% (w/w) chlorhexidine ointment and an intravaginal device with or without progesterone. Chlorhexidine has antibacterial and antifungal properties. This formulation is indicated for topical application for chapped teats, abscesses, and foot rot. Evaluation of its impact as a lubricant was not an objective of this project. A decision was made to include the ointment in the device-placement protocol for 2 reasons. First, the lubricant would facilitate device placement. Second, a 0.1% chlorhexidine acetate solution used in a device-placement protocol reduced the proportion of animals with a vaginal reaction and increased the likelihood of pregnancy at insemination by 9% (18). The impact of chlorhexidine ointment in contact with the vaginal mucosa in dairy cows is unknown. Use of a 1% solution of chlorhexidine gluconate in 2 Welsh ponies was associated with inflammation of the vaginal mucosa but no long-term impact on the health of the animals (19).

Analysis of the hematologic indicators of inflammation in the current study should be interpreted with caution. The University of Guelph’s AHL publishes accepted reference ranges based on 60 clinically healthy mid-lactation cows (12). These reference intervals were used to interpret our results. Treatment with an intravaginal device significantly reduced the circulating leukocyte count, specifically the counts of neutrophils and lymphocytes. The circulating progesterone concentration, a measure of the stage of the estrus cycle, has not been consistently associated with alterations in circulating leukocyte count or leukocyte function (20,21). Absence of leukocytes in vaginal cytologic and histologic samples has been reported during proestrus and estrus (22,23), suggesting a shift in circulating pools of leukocytes toward the vagina in the luteal phase.

The animals enrolled in our study had not displayed signs of estrus before device insertion. Thus, the PRID-treated animals were the only group expected to consistently show a rise in the circulating progesterone concentration at device insertion. Injection of gonadotropin-releasing hormone is known to result in a surge in the production of luteinizing hormone in all cows, ovulation occurring in a proportion of the cows that have a large follicle on the ovary at the time of injection (2). The animals treated with the PID or PRID in our study experienced similar reductions in the circulating leukocyte, neutrophil, and lymphocyte concentrations. The presence of small amounts of cloudy mucus at vaginoscopic examination suggests that all animals reacted to the intravaginal device, the antibacterial lubricant, or both. Neutrophils constitute the 1st line of defense against foreign material in the uterus and presumably in the vagina as well (24). There is no evidence to support an association of hormonal treatment with an alteration in either neutrophil chemotactic ability or phagocytic ability (21).

The biologic significance of small numerical differences in circulating concentrations of leukocytes, specifically neutrophils and lymphocytes, between the device-treated animals and the negative-control animals is not apparent. Endogenous progesterone is thought to increase the susceptibility of the bovine uterus to bacterial infection. In beef cows, the presence of 2 intravaginal devices, irrespective of any drug loading, delayed the resolution of postpartum bacterial colonization of the uterus and vagina (25). Similarly, early resumption of ovarian activity has been associated with increased risk of pyometra (26). However, the impact of the progesterone concentration on the risk of vaginitis in cows with no evidence of pre-existing uterine infection has not been characterized (27).

Data confirm that progesterone supplementation significantly improves the reproductive performance of animals with anestrus and cystic ovarian disease independent of vaginal reaction (6,28). In cows diagnosed as not pregnant 30 to 60 d after insemination, the PRID, used in conjunction with an ovulation-synchronization protocol, improved the probability of pregnancy after fixed-time insemination by 11 percentage points (33.7% to 44.7%) relative to PID treatment (16). Similarly, the probability of pregnancy after 1st insemination performed at observed estrus and time to pregnancy were significantly improved in previously anestrus cows treated with a PRID or a PID for 7 d and then given an injection of prostaglandin (17). In both investigations, the vaginitis score was not associated with the probability of pregnancy after 1st insemination or time to pregnancy.

Increases in the circulating concentration of acute-phase proteins, such as haptoglobin, have been consistently associated with trauma, inflammation, and infection: the concentration of these proteins rises quickly in response to insult and remains elevated until the insult has resolved (29). In the current investigation, the presence of an intravaginal device did not generate a response measurable 7 d after device insertion.

Bacterial culture of swabs of the vagina after a 7-d treatment period revealed moderate growth of coliforms, environmental streptococci and staphylococci, and other gram-positive, rod-shaped organisms. These organisms are considered the normal flora of the skin and the feces of cattle (30). The swab was taken aseptically from the dorsal wall of the caudal vagina, approximately 10 cm rostral to the vulvar lips. Some contamination could have occurred during sample collection. Collection technique has been identified as a potential source of variation (25). Bulman et al (30) sampled close to the cervical os after 14 d of PRID treatment and cultured enterococci, Escherichia coli, and Proteus sp. from 90% of the treated animals and 40% of animals examined with a speculum. Their results are similar to ours despite the difference in sampling location. From repeated sampling, Bulman et al (30) reported no bacterial growth in 90% of the cows 7 d after device removal and in 100% of the cows 14 d after device removal.

Intravaginal device placement provides opportunity for bacterial vaginitis. Assessing the impact of this treatment is difficult. Independent of the risk of introducing bacteria into the vagina, progesterone supplementation has repeatedly been shown to have a positive impact on pregnancy likelihood at insemination (4,31,32). The relative impact of bacterial vaginitis on pregnancy likelihood has not been investigated. Schindler et al (33) considered the impact of all reproductive abnormalities, including metritis, endometritis, and vaginitis and reported delayed onset of cyclicity. However, the relative contribution of impaired genital health on future reproductive performance was small. Both clinical and subclinical endometritis have been associated with increased time to conception (34,35), but a link between bacterial vaginitis and ascending infection of the uterus has not been supported. Sheldon et al (36) introduced bacteria into the vagina of 35 postpartum animals and sampled the uterine endometrium 1 wk later. In all cows, the sampling rod was transferred through the cervix without complication, suggesting that opportunity existed for ascending infection. There was no change in uterine horn involution or bacterial culture results in treated or control animals.

In conclusion, despite slight alterations in the number of circulating leukocytes, neutrophils, and lymphocytes, there was no evidence of a pathologic response to the intravaginal device. All mean hematologic values remained within published reference intervals. The nonspecific inflammatory system, as measured by the acute-phase protein haptoglobin, did not respond to device treatment. Although there was a mild reaction to the presence of the device, as evidenced by slight amounts of purulent mucus on the device, progesterone appeared to reduce the reaction. In addition, the device did not damage the vaginal mucosa.