Vaginal impedometry for detection of optimal breeding time in bitches

Lindsay H. Bergeron Department of Biomedical Sciences, Ontario Veterinary College, University of Guelph, Guelph, ON N1G 2W1, Canada.

Search for other papers by Lindsay H. Bergeron in
Current site
Google Scholar
PubMed
Close
 MSc
,
Stephanie G. Nykamp Department of Clinical Studies, Ontario Veterinary College, University of Guelph, Guelph, ON N1G 2W1, Canada.

Search for other papers by Stephanie G. Nykamp in
Current site
Google Scholar
PubMed
Close
 DVM
,
Brigitte A. Brisson Department of Clinical Studies, Ontario Veterinary College, University of Guelph, Guelph, ON N1G 2W1, Canada.

Search for other papers by Brigitte A. Brisson in
Current site
Google Scholar
PubMed
Close
 DVM, DVSc
,
Pavneesh Madan Department of Biomedical Sciences, Ontario Veterinary College, University of Guelph, Guelph, ON N1G 2W1, Canada.

Search for other papers by Pavneesh Madan in
Current site
Google Scholar
PubMed
Close
 BVSc&AH, MVSc, PhD
,
William Sears Department of Population Medicine, Ontario Veterinary College, University of Guelph, Guelph, ON N1G 2W1, Canada.

Search for other papers by William Sears in
Current site
Google Scholar
PubMed
Close
 MS, MSc
, and
Cathy J. Gartley Department of Population Medicine, Ontario Veterinary College, University of Guelph, Guelph, ON N1G 2W1, Canada.

Search for other papers by Cathy J. Gartley in
Current site
Google Scholar
PubMed
Close
 DVM, DVSc

Abstract

Objective—To compare the efficacy of canine vaginal impedometry in identifying the preovulatory luteinizing hormone (LH) peak to that of currently used methods (serum progesterone concentration measurement, vaginal cytologic evaluation, and vaginoscopy).

Design—Prospective study.

Animals—12 sexually intact female dogs.

Procedures—12 mature postpubertal Beagle (n = 3), Beagle-cross (2), and hound-cross (7) bitches ranging from 7.5 to 27.5 kg (16.5 to 60.6 lb) were enrolled in the study. After the onset of spontaneous proestrus, determined on the basis of appearance of serosanguineous vaginal discharge, serum progesterone assays, vaginoscopy, vaginal cytologic evaluation, and vaginal impedometry were performed daily until approximately 4 days after peak LH concentration (day 0) as measured by radioimmunoassay. Vaginal impedometry was compared against serum progesterone concentration measurement, vaginal cytologic evaluation, and vaginoscopy as a method for accurately identifying the LH peak and therefore the optimal breeding time. Ten of 12 bitches were bred with subsequent assessment of embryos.

Results—Vaginal impedometry accurately predicted the preovulatory LH peak in 5 of 11 bitches. One bitch was removed from the study because data were not collected. Of the remaining 11 bitches, 6 had their LH peak on the day serum progesterone concentration first exceeded 2 ng/mL. Crenulation scores reached 1 (mean, 1.3; 95% confidence interval, 0.8 to 1.7) on day 0 as expected; however, these scores were not significantly different from those on days −1 or 1. Vaginal epithelial cell populations did not change noticeably on day 0. Nine of the 10 bitches that were bred produced viable embryos.

Conclusions and Clinical Relevance—Results suggested that daily use of vaginal impedometry in bitches was unreliable as a method for monitoring periovulatory events. All techniques evaluated (ie vaginal impedometry, serum progesterone concentration assays, vaginoscopy and vaginal cytologic evaluation) frequently produced inaccurate results when used individually. Multiple methods should be used to identify optimal breeding time in dogs.

Abstract

Objective—To compare the efficacy of canine vaginal impedometry in identifying the preovulatory luteinizing hormone (LH) peak to that of currently used methods (serum progesterone concentration measurement, vaginal cytologic evaluation, and vaginoscopy).

Design—Prospective study.

Animals—12 sexually intact female dogs.

Procedures—12 mature postpubertal Beagle (n = 3), Beagle-cross (2), and hound-cross (7) bitches ranging from 7.5 to 27.5 kg (16.5 to 60.6 lb) were enrolled in the study. After the onset of spontaneous proestrus, determined on the basis of appearance of serosanguineous vaginal discharge, serum progesterone assays, vaginoscopy, vaginal cytologic evaluation, and vaginal impedometry were performed daily until approximately 4 days after peak LH concentration (day 0) as measured by radioimmunoassay. Vaginal impedometry was compared against serum progesterone concentration measurement, vaginal cytologic evaluation, and vaginoscopy as a method for accurately identifying the LH peak and therefore the optimal breeding time. Ten of 12 bitches were bred with subsequent assessment of embryos.

Results—Vaginal impedometry accurately predicted the preovulatory LH peak in 5 of 11 bitches. One bitch was removed from the study because data were not collected. Of the remaining 11 bitches, 6 had their LH peak on the day serum progesterone concentration first exceeded 2 ng/mL. Crenulation scores reached 1 (mean, 1.3; 95% confidence interval, 0.8 to 1.7) on day 0 as expected; however, these scores were not significantly different from those on days −1 or 1. Vaginal epithelial cell populations did not change noticeably on day 0. Nine of the 10 bitches that were bred produced viable embryos.

Conclusions and Clinical Relevance—Results suggested that daily use of vaginal impedometry in bitches was unreliable as a method for monitoring periovulatory events. All techniques evaluated (ie vaginal impedometry, serum progesterone concentration assays, vaginoscopy and vaginal cytologic evaluation) frequently produced inaccurate results when used individually. Multiple methods should be used to identify optimal breeding time in dogs.

Vaginal mucus impedance, the electrical resistance of vaginal mucus, may be closely correlated with serum concentrations of progesterone and estradiol.1–3 Opinions differ as to whether monitoring VMI by means of vaginal impedometry could provide an accurate and specific indication of the optimal time to breed in bitches1 or whether this method is only useful for identification of the stage of the estrous cycle of a bitch.2 In sheep,3 cattle,4 pigs,5 and monkeys,6 vaginal impedometry has been used to successfully determine the time of ovulation, as indicated by low electrical resistance during optimal breeding periods. However, bitches have been found to have high electrical resistance during optimal breeding periods, a difference which may be attributed to their unique serosanguineous discharge.1,2 Additionally, bitches range widely in size and a single impedometer is unlikely to reach a consistent location among all breeds, which may make measurement of VMI inconsistent and vaginal impedometry unreliable as a method for daily monitoring of periovulatory events in this species.5,7

Currently used methods of monitoring periovulatory events to identify the optimal time to breed in dogs include serum or plasma progesterone assays, vaginal cytologic evaluation, and vaginoscopy. A serum progesterone concentration > 2 ng/mL is typically accepted as an indicator of the LH surge,8–11 yet a wide range (3 to 10 ng/mL) of progesterone concentrations has been reported as indicative of ovulation.12 These concentrations can be variable among bitches, laboratories, type of assay used, and methods of specimen handling, and may depend on whether progesterone is measured in plasma or serum.8,9,12–18

Vaginal cytologic evaluation allows the exfoliated cells of the vaginal epithelium to be examined. Cell types change throughout the estrous cycle, providing an indication of the cycle stage.19 Vaginoscopy assesses the changing appearance of the vaginal mucosa throughout the estrous cycle (crenulation) and can also provide an indication of the stage of the estrous cycle.20,21 Changes in both vaginal mucosa and epithelium are influenced primarily by circulating estradiol concentrations,8,10,11,22 which can be variable within and among bitches.10,23 In sheep, progesterone concentrations primarily influence VMI, except when progesterone concentrations are low, at which point the estradiol-to-progesterone concentration ratio also influences VMI.3 If this is also true in bitches, vaginal impedometry could provide a more consistent method for monitoring periovulatory events than methods influenced primarily by estradiol, such as vaginal cytologic evaluation and vaginoscopy.

The optimal time to breed is typically 4 to 7 days after the LH surge,11 which requires precise monitoring. Both of the vaginal impedometry studies1,2 previously performed in bitches have relied on surrogate indicators of the LH surge or time of ovulation (eg, serum progesterone assays) to evaluate the efficacy of this technique for monitoring periovulatory events. Critical assessment of these monitoring methods by comparison with direct measurements of LH concentration may enable improvements in breeding management and use of assisted reproductive technologies. Therefore, the objective of the study reported here was to compare the efficacy of canine vaginal impedometry in identifying the preovulatory LH peak against that of currently used methods, including serum progesterone concentration measurement, vaginal cytologic evaluation, and vaginoscopy. We hypothesized that use of vaginal impedometry in conjunction with these other methods would be more accurate than vaginal impedometry alone.

Materials and Methods

Experimental animals—All experimental procedures were performed in accordance with the guidelines established by the Canadian Council on Animal Care and were approved by the Institutional Animal Care and Use Committee at the University of Guelph. Twelve mature postpubertal bitches were enrolled in the study. Bitches were Beagles (n = 3), Beagle crosses (2), or hound crosses (7) with a median weight of 21 kg (46.3 lb; range, 7.5 to 27.5 kg [16.5 to 60.6 lb]). After the onset of spontaneous proestrus, identified on the basis of appearance of serosanguineous vaginal discharge, all animals were housed at the Central Animal Facilitya for the duration of the study.

Vaginal impedometry—Vaginal impedometry was performed with a 2-electrode, rigid, battery-operated impedometer.b Vaginal mucus impedance was measured and recorded daily. The vaginal impedometer was completely inserted into the vagina and spun in one direction to ensure that the electrodes came into full contact with fresh mucus. The impedometer deviceb displays a value that is a mean calculated on the basis of several readings. Those raw values are not available to the operator. Thus the displayed mean VMI was recorded. This process was repeated, ensuring the impedometer was first turned to expose the electrodes to new mucus. After each daily use and also in between bitches being examined on the same day, the impedometer was disinfected with 70% isopropyl alcohol,c followed by concentrated algaecided and then rinsed with distilled water. Vaginal impedometry data for 1 of the 12 bitches were not included because, after the data from this animal were obtained, it was determined that VMI had not been measured on the correct dates corresponding to the LH peak. As such, all VMI data for this bitch were excluded from the analyses.

Daily VMI readings were used for statistical analyses.e In addition, VMI readings for each bitch were plotted in a spreadsheet software programf to allow observation of graphical trends among days from the LH peak. Peak VMI was defined as the highest VMI among days examined. Peak resistance was expected to occur 2 or 3 days after the LH peak.24

Hormonal assays—As previously described,25 blood samples were collected once daily for the detection of the onset of proestrus until approximately 4 days after the LH peak. Serum was assayed by means of a CLIA for progesterone concentration on the day of collection (for the first day of examination and then every second day) at the Animal Health Laboratory.a The remaining serum was stored at −20°C until used for radioimmunoassay. Daily serum progesterone and LH concentrations were later determined by means of a radioimmunoassay validated for use in canine species at the Animal Health Diagnostic Center.g Progesterone concentrations assays were retrospectively found to have been measured on the day of the LH peak for 6 of 12 bitches.

The LH peak (day 0; as measured by radioimmunoassay) was used as a reference, as previously described,25 to determine whether vaginal impedometry, progesterone assays, vaginoscopy, and vaginal cytologic evaluation were accurate in identifying the optimal time to breed in bitches.

Vaginoscopy—Vaginoscopy was performed daily with a pediatric proctoscope.26 The proctoscope was stored in didecyl dimethyl ammonium chloride and N-alkyl dimethyl benzyl ammonium chloride.h Before use, the proctoscope was rinsed with tap water followed by irrigation with sterile saline (0.9% NaCl) solution.i Vaginal crenulation was assessed in the cranial aspect of the vagina with the dorsal caudal median fold in view. Descriptions of the changing appearance of vaginal mucosa throughout the estrous cycle20,21 were assigned a crenulation score on a scale from 0 to 3: 0 = pale red edematous mucosa (expected prior to the preovulatory LH peak), 1 = shrinking pale mucosa without angulation (expected at the preovulatory LH peak), 2 = shrinking pale mucosa with further angulation (expected at ovulation), or 3 = maximal wrinkling and pale mucosa (expected at oocyte maturation and maximum fertility). After observing crenulation, cotton-tipped swabsj moistened with tap water were inserted through the proctoscope and rotated in one direction to obtain cells for vaginal cytologic evaluation.

Vaginal cytologic evaluation—Vaginal cytologic slides were collected and evaluated as previously described.22 Slides were stained with a modified Wright-Giesma stain.k A slide characteristic of estrus had 100% superficial cells, > 80% pyknotic or absent nuclei, no polymorphonuclear cells, and little or no background debris.22

Artificial insemination—As a component of a separate research experiment (unpublished data), 10 of the 12 bitches were artificially bred with fresh semen. Optimal breeding time was identified on the basis of results of progesterone assays, vaginoscopy, and vaginal cytologic evaluation. Eleven to 15 days after the LH peak (as measured by radioimmunoassay), ovariohysterectomy was performed. For the 10 bitches that were bred, the uterine horns were flushed for embryos and the collected embryos were visually assessed.

Statistical analysis—Vaginal impedometry analyses were performed in 2 ways. A general linear mixed model was fitted.e The 2 daily VMIs for each bitch were input into the statistical software,e and mean daily VMIs were used for analysis. The fixed effect was the day of the LH peak. The model was simplified by removing nonsignificant terms. Because repeated measurements were performed in the animals, a potential autocorrelation needed to be accounted for. The following correlation error structures, offered by the statistical software,e were attempted, and the Akaike information criterion was used to select one of the following: autoregressive 1, autoregressive heterogenous 1, toeplitz and banded toeplitz 2 through 7 as well as the heterogenous versions, and so-called unstructured and the banded versions unstructured 2 through 7. Finally, a simple random effect of dog was treated as a blocking variable.

To examine the ANOVA assumptions, thorough residual analyses were conducted. This included formally testing the residuals for normality by means of Kolmogorov-Smirnov, Cramér-von Mises, Shapiro-Wilk, and Anderson-Darling tests in statistical softwaree and plotting the residuals against the predicted values and explanatory variables used in the model. Such analyses may reveal outliers, unequal variance, or other problems that need to be addressed, and may suggest the need for data transformation. Least square means were calculated for each day from the preovulatory LH peak and the differences (for untransformed data) or ratios (for transformed data) of least square means were used to determine significant (P < 0.05) differences between days, by use of the Tukey-Kramer method.

As per the impedometer user manual,24 if bitches had VMI peaks occurring on days 2 or 3 after the LH peak, this technique was categorized as efficacious for monitoring periovulatory events and identifying the optimal time to breed in bitches. The mean and 95% CI for these vaginal impedometry results were determined with exact binomial CIs as by the Sterne method.27

Additionally, the mean ± SD progesterone concentrations on day 0 and 2 as determined by means of radioimmunoassay (n = 12) and CLIA (6 [CLIA progesterone assays were retrospectively found to have been measured on the day of the LH peak for 6 of 12 bitches]) were calculated with spreadsheet softwaref to compare hormonal concentrations on the days corresponding to the LH peak and ovulation, respectively, on the basis of results of the 2 assays.

For vaginoscopy, a general linear mixed model was fitted.e Daily crenulation scores for each bitch were input and used for analysis. The fixed effect was the day of the LH peak. The model was simplified by removing nonsignificant terms. Because variables were repeatedly measured in the same dogs, a potential autocorrelation needed to be accounted for. Correlation error structures were attempted, as described for vaginal impedometry. Residual analyses, also as described for vaginal impedometry, were conducted to examine the ANOVA assumptions. Least square means were calculated for each day from the preovulatory LH peak and the differences (for untransformed data) or ratios (for transformed data) of least square means were used to determine significant (P < 0.05) differences between days, by means of the Tukey-Kramer method. For vaginal cytologic evaluation, median and 95% CI were calculated by means of the median unbiased estimate method.

Results

Vaginal impedometry—For the general linear mixed model analysis, VMIs were log transformed, resulting in all assumptions of the model being met. The random error structure produced the lowest Akaike information criterion (Table 1). Briefly, daily median VMI ranged from 281 (day −2) to 562 (day 2). Significant increases in median VMI measurements were observed among several days. The highest median VMIs were reached on day 2; however, median VMI on this day was not significantly different from median VMIs on previous or subsequent days.

Table 1—

Daily VMI readings in Beagle (n = 3), Beagle-cross (2), and hound-cross (7) bitches with weight ranging from 7.5 to 27.5 kg (16.5 to 60.6 lb).

DayMedian VMI95% CI
−3341a,b230 to 505
−2281a193 to 412
−1405a,b275 to 595
0536b364 to 790
1449a,b305 to 661
2562b382 to 827
3491b,c334 to 723
4341a,b230 to 505

Day 0 represents the preovulatory LH peak. All values reported have been back transformed from the log scale.

Values without a common superscript letter differ significantly (P < 0.05).

From the analysis with exact binomial CIs (Sterne method27), our data suggested that vaginal impedometry was an accurate method of monitoring periovulatory events in only 5 of 11 bitches, or 45% of animals (with Sterne limits of 20% and 74%; Figure 1). Bitches that did not have VMI peak at the correct time had the peak occur too early, did not have a peak at all, or had a large false peak (Figure 2).

Figure 1—
Figure 1—

Example of mean VMI relative to days from the preovulatory LH peak for a single representative bitch for which VMI peaked on an appropriate day (ie, 2 or 3 days after the LH peak). Mean VMI is marked (squares). The days when peak VMI should occur are circled with dashed lines, and the days of optimal breeding are circled with solid lines. Day 0 represents the LH peak. Lutenizing hormone concentrations are marked with circles. For a single reading, the impedometer deviceb displays a mean value for the animal, calculated on the basis of several readings.

Citation: Journal of the American Veterinary Medical Association 245, 12; 10.2460/javma.245.12.1360

Figure 2—
Figure 2—

Example of mean VMI relative to days from the preovulatory LH peak for a single representative bitch for which the VMI did not peak on the appropriate day (ie, 2 or 3 days after the LH peak). A false peak in VMI can be seen on day 0, which was larger than the subsequent spike in VMI on day 3. See Figure 1 for remainder of key.

Citation: Journal of the American Veterinary Medical Association 245, 12; 10.2460/javma.245.12.1360

Hormonal assays—Each bitch was found to have anticipated patterns of hormone secretions, as previously reported.25 Mean ± SD progesterone concentration (as measured by means of radioimmunoassay) on day 0 was 3.07 ± 1.06 ng/mL. On day 2, the day of presumed ovulation, mean ± SD progesterone concentration was 6.01 ± 2.11 ng/mL, as previously reported.25 Mean ± SD CLIA progesterone concentration was 3.90 ± 1.56 ng/mL on day 0 and 5.60 ± 2.28 ng/mL on day 2. Only 6 of 11 bitches had their LH peak (day 0) occur on the day radioimmunoassay progesterone concentrations first exceeded 2 ng/mL. This analysis could not be performed for CLIA progesterone concentrations because these were not performed daily. On the basis of an LH concentration cutoff of 1 ng/mL to identify the LH surge and the highest assayed LH concentration to identify the day of the LH peak, 10 of 12 bitches had their LH peak and surge on the same day and 2 of 12 bitches surged the day before their peak. Taking doubling of radioimmunoassay serum progesterone from one day to the next as an indication of LH surge,15 only 6 of 12 bitches would have had their LH surge properly identified. Progesterone concentration doubled in 2 bitches the day after the LH peak (or surge); in 2 bitches, progesterone concentration never doubled from one day to the next during the course of the study (up to 4 days past the LH peak).

Vaginoscopy—For the general linear mixed model analysis, all assumptions of the model were met and transformation of the data was not necessary. The random error structure produced the lowest Akaike information criterion (Table 2). Briefly, daily mean crenulation scores ranged from 0.1 (95% CI, 0+ to 0.6) on day −3 to 3.0 (95% CI, 2.5 to 3–) on day 4. There were significant increases in mean crenulation scores observed among days, which progressively increased from day −3 to 4. Crenulation scores reached 1 (mean, 1.3; 95% CI, 0.8 to 1.7) on the day of the LH peak, as expected; however, scores on day 0 were not significantly different from scores on either day −1 or 1.

Table 2—

Daily crenulation scores evaluated by means of vaginoscopy for the bitches in Table 1.

DayMean crenulation score95% CI
−30.1a,b0+ to 0.6
−20.6b,c0.1 to 1.0
−10.9c,d0.5 to 1.4
01.3d0.8 to 1.7
11.6d,e1.1 to 2.0
22.1e,f1.7 to 2.5
32.6f,g2.2 to 3.1
43.0g2.5 to 3–

Descriptions of the changing appearance of the vaginal mucosa throughout the estrous cycle20,21 were assigned a crenulation score on a scale from 0 to 3; 0+ represents a value that is arbitrarily close to 0, and 3– represents a value that is arbitrarily close to 3.

Values without a common superscript letter differ significantly (P < 0.05).

See Table 1 for remainder of key.

Vaginal cytologic evaluation—All 12 bitches had vaginal cytologic findings characteristic of estrus before the day of their LH peak. The vaginal epithelial cell population viewed on the cytologic slides did not change noticeably at the LH peak.

Artificial insemination—Of the 10 bitches bred, 9 produced viable embryos, confirming the accurate detection of optimal breeding time, on the basis of directly measured LH concentration as a reference.

Discussion

The results of the present study of 12 mature postpubertal bitches suggested that daily vaginal impedometry is unreliable as a method for monitoring of periovulatory events. All techniques evaluated (ie, progesterone assays, vaginoscopy, vaginal cytologic evaluation, and vaginal impedometry) frequently produced inaccurate results when used individually. The impedometer user manual24 states that the profile trace of the VMI readings is important, not the numerical values. The manual also warns users of a false peak, which may occur before and at a lower resistance than the real peak. In this study, a peak was thereby defined as the highest VMI reading retrieved among days examined. Peak resistance occurs 2 or 3 days after the LH peak and the optimal time to breed is 2 days after the peak VMI reading.24 Thus, if bitches had VMI peaks occurring on days 2 or 3 after the LH peak, this technique was categorized as efficacious for monitoring periovulatory events and identifying the optimal time to breed in bitches. Although the highest median VMI was reached on day 2, median VMI on this day was not significantly different from values obtained on previous or subsequent days.

Additionally, in the present study, only 5 of 11 bitches had their peak VMI measurement occur at the appropriate time according to the impedometer user manual24 (Figure 1). Bitches that did not have VMI peak at the correct time had the peak occur too early, did not have a peak at all, or had a large false peak (Figure 2). This false peak would complicate the use of vaginal impedometry for monitoring periovulatory events because even if there were an increase in VMI at the appropriate time (day 2 or 3), if this increase were not larger than the false peak, the user would incorrectly (prematurely) identify the optimal time to breed. Additionally, if this method were being used to determine the optimal time to breed artificially (eg, with fresh semen), there would be little time to prepare for breeding because the user would need to first wait to observe whether VMI reached a peak (ie, wait until VMI declines).

Furthermore, the ability of the impedometerb to measure the resistance of pericervical mucus may be compromised by the length of the device, which is too short to reach the cranial vagina in any bitch larger than a Beagle. However, our data did not suggest a compromised ability to determine the optimal time to breed because the 5 bitches that had their peak VMI occur at the appropriate time were of varying size and weight (7.5 to 27.5 kg [16.5 to 60.6 lb]).

Although the impedometer user manual24 states that the unit of measurement of VMI is unimportant to the analyses, and that VMI is simply reported in arbitrary units, it should be noted that a prior study1 describes trends in VMI in Ohms. Vaginal mucus impedence measurements as previously reported have been shown to reach a plateau lasting 4 to 11 days.1 Because the optimal time to breed has been identified during peak VMI and these peak values reach a plateau, it has been suggested that recording values in 2-day intervals is sufficient to determine the optimal time to breed.1 This plateau in VMI was not observed in the present study, suggesting that daily examinations would be required.

In sheep, VMI is primarily influenced by progesterone concentrations but is also affected by the ratio of estradiol to progesterone concentration when progesterone concentrations are decreasing or are low.3 Vaginal mucus impedance may be controlled by estradiol concentration in bitches, an assumption that has been made based on the parallel changes observed for results of vaginal cytologic evaluation and for vaginal impedance, which are primarily affected by estradiol concentration.1 Because decreased estradiol concentration have been reported to occur near the LH peak,12,15–17 the results of our study, which found no significant relationship between LH and VMI, suggest that vaginal impedance was not controlled by estradiol concentration. It may be that vaginal impedance was controlled primarily by progesterone concentration or, as it is in sheep,3 by the progesterone-to-estradiol concentration ratio. The unique pattern of changing progesterone concentrations in bitches (a result of preovulatory luteinization of follicles) could be responsible for the unique pattern of changes in VMI, compared with other species; however, this would require further investigation.

Although both CLIA and radioimmunoassay progesterone concentrations for all bitches exceeded 2 ng/mL on the day of the LH peak, only 6 of 11 bitches had their LH peak occur on the day radioimmunoassay progesterone concentrations first exceeded 2 ng/mL. It should be noted that although the changes in hormonal concentrations observed among days from the LH peak were as anticipated (on the basis of a previous report25), the values on the days of interest (day 0 and 2) did not allow for precise identification of the LH peak and ovulation. This demonstrated that measurement of progesterone concentration could serve as an estimate of the LH peak but should not be used to reliably identify the optimal breeding time when used alone. On the basis of the variability in progesterone concentrations between CLIA and radioimmunoassay, the day that progesterone concentrations first exceeded 2 ng/mL may differ among assays, making identification of the LH peak inconsistent.

Measurement of progesterone concentration is often used as an indirect indicator of the LH surge because it is more economical to assay.10 Our data support previous speculation that sole use of progesterone concentration cutpoints to predict the LH surge may be unreliable,28 in that we could not reliably use traditional methods of cutoffs or doubling of progesterone concentration to correctly identify the LH surge. This could be for a number of reasons, many of which have been previously documented, including variability among bitches, laboratories, the type of assay used, specimen handling, and whether plasma or serum is used.8,9,12–18 Progesterone assays should certainly be used in ovulation monitoring but in tandem with other measures, perhaps including serial ovarian ultrasoonography25 or in-house LH assays.

A progressive increase in crenulation scores relative to days from the preovulatory LH peak was observed in this study, which made it difficult to know when specific periovulatory events had occurred. The crenulation score did reach 1 (mean, 1.3; 95% CI, 0.8 to 1.7) on the day of the LH peak, as expected. However, the crenulation score on day 0 was not significantly different from the crenulation scores on either day −1 or 1, making the precise detection of the LH peak unlikely. Crenulation is influenced by declining estradiol concentrations and is subjective in interpretation of the degree of wrinkling.20 These characteristics make evaluation of crenulation a variable and unreliable method of monitoring periovulatory events when performed alone, as has been previously reported.29

Vaginal cytologic evaluation was ideal for determining what stage of the estrous cycle a bitch was in, specifically when a bitch was entering into estrus. However, there was no noticeable change in the population of vaginal epithelial cells at the LH peak or at presumed ovulation, making this technique unsuitable for accurately identifying the optimal time to breed in bitches if used alone.

In this study, we experienced the same difficulties as have been previously documented with vaginal cytologic evaluation.11,19,29 For example, variability in the amount and time maximum cornification is visible in a vaginal smear varies among bitches, making it difficult to know the exact time for optimal breeding.11 Error has also been documented when the appearance of anuclear cells is relied on as an indication of the progression from proestrus to estrus. In addition, the type of stain used can introduce variability in determining whether a cell appears as anuclear or not.19 Vaginal cytologic evaluation is ultimately a retrospective method of monitoring periovulatory events in bitches.29 Monitoring the progression from estrus to diestrus (when superficial cells begin to decrease) can allow for retrospective confirmation of whether breeding was performed at the optimal time.1,11 This method has been shown to be reliable (if performed frequently) to determine the stage of the estrous cycle12 but not to precisely or prospectively determine the optimal time to breed.

It should also be noted that measurements of VMI and crenulation scores were both acquired easily. However, measurement of progesterone concentrations first required serum to be separated from blood samples and then assayed. Use of the impedometer was also less tedious than creating and evaluating a vaginal cytologic slide, which used disposable swabs to retrieve cells, which were then transferred onto a slide and stained, followed by microscopic evaluation.30 Both the impedometer (used for vaginal impedometry) and the pediatric proctoscope (used for both vaginal cytologic evaluation and vaginoscopy) required thorough disinfection after each use. Disinfection is a concern when these pieces of equipment are used because they have the potential to spread infectious agents. Vaginal impedometry and progesterone assays produce raw values as assessed by a device, whereas vaginal cytologic evaluation and vaginoscopy both require visualization and interpretation by someone knowledgeable and experienced. This subjective interpretation may be inconsistent among users and among days with the same user.30 All techniques required serial monitoring for optimal reliability.

Thus, it is recommended that the accuracy of identifying optimal time to breed, level of invasiveness, and ease of use of each technique, in addition to available resources, be considered when determining monitoring methods. Vaginal impedometry, progesterone assays, vaginal cytologic evaluation, and vaginoscopy all have advantages and disadvantages for use in determining the optimal time to breed in bitches. The results of the present study suggested that vaginal impedometry would not be accurate when used alone to identify optimal breeding time and did not improve accuracy even when used in combination with the other currently used methods. By critically assessing these monitoring methods on the basis of quantitative measurement of LH concentration as a reference, they can be accurately reviewed, which may allow for improvements in breeding strategies and improved success of assisted reproductive technologies in dogs.

ABBREVIATIONS

CI

Confidence interval

CLIA

Chemiluminescence immunoassay

LH

Luteinizing hormone

VMI

Vaginal mucus impedance

a.

University of Guelph, Guelph, ON, Canada.

b.

Draminski Electronic Oestrous Detector, Firma Draminski, Olsztyn, Poland.

c.

Commercial Alcohols, Brampton, ON, Canada.

d.

Quat-40, Jes-Chem Ltd, Guelph, ON, Canada.

e.

Proc MIXED, SAS, version 9.3, SAS Institute Inc, Cary, NC.

f.

Excel, Microsoft Corp, Redmond, Wash.

g.

Cornell University, Ithaca, NY.

h.

Germiphene 1:64, Ontario Veterinary College, Guelph, ON, Canada.

i.

Baxter Corp, Toronto, ON, Canada.

j.

Kalayjian Industries Inc, Signal Hill, Calif.

k.

Protocol Hema-3, Fisher Scientific Co LLC, Kalamazoo, Mich.

References

  • 1. Günzel AR, Koivisto P, Fougner JA. Electrical resistance of vaginal secretions in the bitch. Theriogenology 1986; 25: 559570.

  • 2. Leidl W, Stolla R. Measurement of electric resistance of the vaginal mucus as an aid for heat detection. Theriogenology 1976; 6: 237249.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 3. Bartlewski PM, Beard AP, Rawlings NC. The relationship between vaginal impedance and serum concentrations of estradiol and progesterone throughout the sheep estrous cycle. Theriogenology 1999; 51: 813827.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 4. Edwards DF. Proposed instrumentation to determine the optimum time to inseminate cattle by measurement of vaginal impedance. Med Biol Eng Comput 1980; 18: 7380.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 5. Rezác P, Olic I. Relationship between opposite changes of vaginal and vestibular impedance during estrous cycle in sows. Theriogenology 2006; 66: 868876.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 6. Fischer L, Germain G, Florence G, et al. Changes in electricalimpedance of the vaginal medium during the menstrual-cycle of female Rhesus-monkeys (Macaca mulatta). J Med Primatol 1990; 19: 573582.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 7. Rezác P, Poschl M, Krivanek I. Effect of probe location on changes in vaginal electrical impedance during the porcine estrous cycle. Theriogenology 2003; 59: 13251334.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 8. Mestre J, Wanke M, Sucheyre S. Exfoliative vaginal cytology and plasma concentrations of progesterone, luteinizing hormone and oestradiol-17β during oestrus in the bitch. J Small Anim Pract 1990; 31: 568570.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 9. Wright PJ. Application of vaginal cytology and plasma progesterone determinations to the management of reproduction in the bitch. J Small Anim Pract 1990; 31: 335340.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 10. Hewitt D, England G. Assessment of optimal mating time in the bitch. In Pract 2000; 22: 2433.

  • 11. England GCW, Concannon PW. Determination of the optimal breeding time in the bitch: basic considerations. In: Recent advances in small animal reproduction. Document No. A1231.0602. Ithaca, NY: International Veterinary Information Service, 2002.

    • Search Google Scholar
    • Export Citation
  • 12. Wright PJ. Practical aspects of the estimation of the time of ovulation and of insemination in the bitch. Aust Vet J 1991; 68: 1013.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 13. Hase M, Hori T, Kawakami E, et al. Plasma LH and progesterone levels before and after ovulation and observation of ovarian follicles by ultrasonographic diagnosis system in dogs. J Vet Med Sci 2000; 62: 243248.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 14. Volkmann DH. The effects of storage time and temperature and anticoagulant on laboratory measurements of canine blood progesterone concentrations. Theriogenology 2006; 66: 15831586.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 15. Concannon PW, Hansel W, Visek WJ. The ovarian cycle of the bitch: plasma estrogen, LH and progesterone. Biol Reprod 1975; 13: 112121.

  • 16. Concannon PW, Hansel W, McEntee K. Changes in LH, progesterone and sexual behavior associated with preovulatory luteinization in the bitch. Biol Reprod 1977; 17: 604613.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 17. Concannon PW, McCann JP, Temple M. Biology and endocrinology of ovulation, pregnancy and parturition in the dog. J Reprod Fertil Suppl 1989; 39: 325.

    • Search Google Scholar
    • Export Citation
  • 18. Hayer P, Gunzel Apel AR, Luerssen D, et al. Ultrasonographic monitoring of follicular development, ovulation and the early luteal phase in the bitch. J Reprod Fertil Suppl 1993; 47: 93100.

    • Search Google Scholar
    • Export Citation
  • 19. Johnston SD, Root Kustritz MV, Olson PN. Vaginal cytology. In: Johnston SD, Root Kustritz MV, Olson PN, eds. Canine and feline theriogenology. Philadelphia: WB Saunders Co, 2001;33.

    • Search Google Scholar
    • Export Citation
  • 20. Okkens AC, Bevers MM, Dieleman SJ, et al. Fertlity problems in the bitch. Anim Reprod Sci 1992; 28: 379387.

  • 21. Lulich JP. Endoscopic vaginoscopy in the dog. Theriogenology 2006; 66: 588591.

  • 22. Olson PN, Thrall MA, Wykes PM, et al. Vaginal cytology. I. A useful tool for staging the canine estrous cycle. Compend Contin Educ Pract Vet 1984; 6: 288297.

    • Search Google Scholar
    • Export Citation
  • 23. Hadley JC. Total unconjugated oestrogen and progesterone concentrations in peripheral blood during the oestrous cycle of the dog. J Reprod Fertil 1975; 44: 445451.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 24. Draminski electronic oestrous detector user manual. Olsztyn, Poland: Firma Draminski, 2008.

  • 25. Bergeron LH, Nykamp SG, Brisson BA, et al. An evaluation of B-mode and color Doppler ultrasonography for detecting periovulatory events in the bitch. Theriogenology 2013; 79: 274283.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 26. Lindsay FEF. The normal endoscopic appearance of the caudal reproductive tract of the cyclic and non-cyclic bitch: post-uterine endoscopy. J Small Anim Pract 1983; 24: 115.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 27. Santner TJ, Duffy DR. The statistical analysis of discrete data. New York: Springer-Verlag, 1989; 3343.

  • 28. Fontbonne A, Malandain E. Ovarian ultrasonography and follow-up of estrus in the bitch and queen. Waltham Focus 2006; 16: 2229.

  • 29. Levy X, Fontbonne A. Determining the optimal time of mating in bitches: particularities. Rev Bras Reprod Anim 2007; 31: 128134.

  • 30. Moxon R, Copley D, England GCW. Quality assurance of canine vaginal cytology: a preliminary study. Theriogenology 2010; 74: 479485.

    • Crossref
    • Search Google Scholar
    • Export Citation
All Time Past Year Past 30 Days
Abstract Views 163 0 0
Full Text Views 1215 950 43
PDF Downloads 333 178 14
Advertisement