Murine ovulatory response to ultrasound exposure and its gynecological relevance

Ultrasoundin Med. & Biol. Vol. 14, No. 6, pp. 485-491, 1988 Printed in the U,S.A.

0301-5629/88 $3.00+ .00 © 1988 PergamonPressplc

OOriginal Contribution M U R I N E OVULATORY R E S P O N S E TO U L T R A S O U N D E X P O S U R E A N D ITS GYNECOLOGICAL RELEVANCE

ALLEN H. GATES/f'*'II'~f~fEDWIN L. CARSTENSEN,*'** SALLY Z. CHILD,** W . J. HALL¶ a n d CATHERINE L. MACZYNSKI§ Departments of ~Biophysies and §Pediatrics, IIEnvironmental Health Sciences Center, Divisions of ¶Biostatistics and ~f~'Genetics,The University of Rochester School of Medicine and Dentistry, Rochester, NY 14642, U.S.A. and ~:Department of Electrical Engineering, The University of Rochester, Rochester, NY 14627, U.S.A. (Received 19 August 1987; in finalJbrm 18 January 1988)

Abstract--Ultrasonographic assessment of ovarian follicular maturity was reportedly associated with atypically early ovulation in women; related studies reported reduced litter sizes in rats. To confirm these findings, mice which were midway between ovulatory gonadotropin (LH or human chorionic goundotropin) stimulation and ovulation, were sham- or ultrasound-treated periovarially for 5 min. Exposure was at a spatial average, temporal maximum intensity of 60 W/cm 2. Carrier frequency in the pulse was 2.2 MHz, pulse length was 10 ;ts, and pulse repetition frequency was 200 Hz. Spatial average, temporal average intensity was 0.12 W/cm 2. At autopsy, ultrasound- and sham-treated groups responded similarly in proportions ovulating and in mean ova ovulated. Combined experiments had a 97% chance of detecting a significant (>1 h) advance in ovulation time, had it occurred. Thus, our adequately sensitive mammalian ovulatory tests revealed no association of ultrasound with decrease in ovum number or acceleration in ovulation time (as reported in humans).

Key Words: Ultrasound, Ovulation timing, Ovum, Mouse, Artificial insemination, Fertilization in vitro, Graafian follicle, Gynecology, Human chorionic gonadotropin, Infertifity, Luteinizing hormone, Meiosis, Ovary, Ovulation induction, Ultrasonic tissue effects.

I. INTRODUCTION

Echographic monitoring of h u m a n follicle growth was reported to be associated with atypically early ovulation in a retrospective clinical study (Testart et al., 1982). Patients exposed to ultrasonography at varying times during the 36 h after ovulatory gonadotropin stimulation were claimed to have ovulated prematurely with significantly greater frequency than unexposed patients. Premature ovulation, if induced by ultrasound, could release ova which had undergone altered rates of meiotic division; resultant offspring could manifest increased lethality, malformations, or chromosomal anomalies (BomselHelmreich, 1976, 1985). Any ultrasound-induced, premature ovulation could be a major concern to the gynecologist, especially if it should occur in conjunction with clinical regimens intended to alleviate infertility. Other deleterious reproductive effects have also been reported following ultrasound exposures during periovulatory periods. Demoulin et al. (1985) found such exposures in women were associated with decreased pregnancy rates, and in rats periovulatory ex-

The gynecological application of ultrasonography for Graafian follicle growth monitoring is becoming increasingly frequent. The procedure is used to identify mature follicles either as an aid to retrieval ofoocytes, e.g. for in vitro fertilization, or as a diagnostic tool for predicting time of ovulation, e.g. for artificial insemination (Bomsel-Helmreich, 1985; Hansmann et al., 1986). Relevant to biological safety, obstetrical diagnostic levels of ultrasound have not shown reproducible hazardous effects; this is true for human development (Ziskin, 1986), for animal development (Carstensen and Gates, 1985; O'Brien, 1985), and for genetic factors (Thacker, 1985). However, questions have been raised concerning the possibility that gynecological ultrasound could produce adverse effects upon ovarian function and subsequent fertility.

~"To w h o m correspondence should be sent: Environmental Health Sciences Center, University of Rochester Medical Center, Rochester, NY 14642, U.S.A. 485

486

Ultrasound in Medicineand Biology

posures significantly reduced prenatal litter sizes. Sonication of mouse ovaries (at levels above those characteristic of clinical therapeutic ultrasound) for up to seven days was associated with deleterious histological changes in oocytes and ovaries: vacuolization of the germinal vesicle, dissolution of the nucleolus, severe damage to the zona pellucida and an increased frequency of polyovular follicles--with up to six ova per follicle (Bailey et al., 1983). After ovulation, in vitro exposure of mouse oocytes, followed by in vitro fertilization and embryo transfer, was accompanied by an increased rate of postimplantation mortality (Puissant et al., 1984). However, ovarian sonography is useful as the most proximal predictor of impending ovulation (Campbell, 1985), and a higher fecundability rate was even claimed following such diagnostic use in artificially inseminated patients (Marinho et al., 1982). In order to fully evaluate the net benefit of gynecological diagnostic procedures in which periovulatory oocytes are exposed to ultrasound, further prospective experimental studies are required. The objective of our study was to determine whether preovulatory pulsed ultrasound exposure would affect temporal or quantitative aspects of ovulation using the laboratory mouse as an experimental test system. II. MATERIALS A N D M E T H O D S A. Animals All of the mice used were raised and housed in the Inbred Mouse Unit (Laboratory Animal Services Core) of the Environmental Health Sciences Center, with generally accepted standards of care being provided by the staff of the University's Vivarium. Room temperature was 22 ___3°C, relative humidity was 50 ___ 20%, and lighting (fluorescent) was provided for 14 h daily (see Experimental design). Cages were clear polycarbonate and had filter bonnets (Filtek). The diet was R M H 2000 (Country Foods, Division of Agway, Inc.). The mouse stocks used were inbred BALB/cGa and C 129F~ hybrids from BALB/c dams and strain 129/RrGa sires (see Gates and Bozarth, 1978, for strain origins). B. Procedures 1. Injection materials. Pregnant mare serum go° nadotropin (PMSG, from Sigma Chemical Co.) and human chorionic gonadotropin (hCG, from Ayerst Laboratories, Inc.) were freshly diluted, in chilled 0.85% NaC1, from aliquots of bioassayed solution maintained at - 7 0 ° C . Sodium pentobarbital was from D-M Pharmaceuticals, Division of Lemmon Co., Sellersville, PA 18960.

Volume14, Number 6, 1988 2. Animal ovulation models. Our animal models involved ovulatory gonadotropins corresponding to those in the human study by Testart et al. (1982). These were endogenous luteinizing hormone (LH) and exogenous hCG. In the adult mouse, hCG-stimulation was used because it results in reasonably good timing of ovulation. However, for LH-stimulated ovulation, more predictable mouse responses (ineluding ovulation timing) have been attained using an immature mouse model (Gates and Bozarth, 1978). The ovulation resembles that occurring spontaneously in adults, with the timing being regulated by pituitary release of LH during a critical period on the previous afternoon, i.e. at two days after PMSG injection in this model (Gates, 1969). 3. Ultrasound exposure. The mice were anesthetized by sodium pentobarbital injection (ip, 60 mg/kg), shaved ventrally and mounted by their limbs to a rack which was attached to a three-way positioner. The mice were then partially submerged (posterior to the thorax) in a 37°C water bath for ultrasound exposure. The source was a 2.5-cm diameter unfocused piezoceramic element operating at 2.2 MHz. Source to animal distance was approximately 4 cm. To minimize the effects of variations in the local intensity in the near field, the animals were moved slowly back and forth so that the region encompassing both ovaries was irradiated throughout the 5 rain exposure time. Although somewhat longer than is typical of diagnostic ultrasound, pulse lengths of 10 us were used with the rationale that this would maximize the probability of the occurrence of transient cavitation and still be short enough to avoid the complications of rectified diffusion. Spatial average, temporal average intensity was determined by dividing total acoustic power, as measured with an absorbing radiation force target, by the cross-sectional area of the source. Spatial average, temporal maximum intensity could be determined under these conditions simply by multiplying the temporal average intensity by the duty cycle (1:500, i.e. 200 pulses per second). Determined in this way, the spatial average, temporal average intensity used in these experiments was 0.12 W/cm 2 and the spatial average, temporal maximum intensity was 60 W/cm 2. Because of the nature of the near field, spatial peak intensities were probably twice the spatial average values. For comparison, Stewart (1983) reported that temporal maximum intensities of the devices used in obstetrical practice averaged 30 W/cm z with an upper limit to the range of values in excess of 1000 W/cm 2. In view of both the difference in attenuation between mouse and human exposures and, particularly, the relatively long pulses used in

Murine ovulatory response to ultrasound @A. H. GATESel al. our experiments, we conclude that, from the standpoint of the production o f acoustical cavitation, our exposures were representative of the most extreme conditions which would be found in obstetrical or gynecological practice.

4. Autopsy. The mice were sacrificed (cervical dislocation), their fallopian tubes were removed, and ova were recovered in saline (Gates, 1971). The presence (and number) o f ova that were fresh, i.e. with cumulus cells, indicated the progress o f ovulation. For example, the earliest appearance o f such tubal ova approximated the earliest time o f onset o f ovulation since ova are transported from a ruptured ovarian follicle to the oviduct in a matter o f minutes. The mice had beefi given code numbers at exposure so that the dissecting technicians could not distinguish between ultrasound- and sham-treated animals until after the autopsies were completed and the results were recorded. C. Experimental design Experiment 1. Here a m o u s e model was employed in which ovulation was controlled by exogenous ovulatory hormone, hCG. The timing o f events is given in Table 1 along with that for Experiment 2, for comparison. Strain BALB/c females (5-7 months old) were maintained in a room illuminated from 0500 to 1900 h Standard Time (ST), daily. On experimental Day 0 they were injected ip with 2 IU P M S G and on Day 2, with 2 IU hCG (see Sec. liB.). Seven hours later (+ 1 h) ultrasound or sham treatment was administered. Groups of mice were autopsied hourly from 12-15 h after hCG. This experiment comprised three replicates, with approximately equal numbers of mice in each group in each replicate (see Table 2 for exceptions) and with a total o f 72 mice.

Table 1. Times of diurnal, physiological and procedural events in Experiments l (adult mice) and 2 (prepubertal mice) Event

Day

Experiment 1 STy"and CT$

Middle of dark period PMSG§ injection Ovulatory hormone" Ultrasound exposure T~ autopsy¶

daily 0 2 2 2

2400 1400 0200 0930 1600

Experiment2 ST CT 1000 0900 0330 0930 1400

2400 2300 1730 2330 0400

~-Standard time. ¢ Colony time. § Pregnant mare serum gonadotropin. " For Experiment 1: injection of human chorionic gonadotropin; for Experiment 2: endogenous luteinizing hormone release (cited from Gates, 1969). ¶ The autopsy at which ovulation (presence of > 1 tubal ovum) was first observed.

487

Experiment 2. Here a mouse model was used with ovulation controlled by endogenous ovulatory hormone, i.e. luteinizing h o r m o n e (LH). C 129Fj hybrid offspring were raised from one week o f age under reversed illumination, with the lights off from 0500 to 1500 h ST. At 23 days o f age, on experiment Day 0, females were injected with P M S G ( 1 IU, ip) to stimulate ovarian follicle maturation (Gates and Bozarth, 1978). On Day 2 the mice were exposed as in Experiment 1 at 0930 h ST, at an estimated 6 h after luteinizing h o r m o n e (LH) activation (as d e t e r m i n e d by Gates, 1969). Groups were autopsied hourly at 5-7 h after the ultrasound or sham treatment. This experiment comprised seven replicates, with virtually equal numbers of mice in each group in each replicate (see Table 3 for exceptions) and with a total o f 150 mice. D. Statistical methods Pairs o f percentages were compared by standard normal approximation methods; pairs o f means were compared by t-tests. A more comprehensive model was also used for analyzing the frequencies o f ovulation over time: specifically, a logit model with linear time effects. The log-odds of ovulation at any given time are defined as ln(p/q), where p is the probability of ovulation occurring, and q (= 1 - p) is the probability of ovulation not occurring. With t representing the n u m b e r of hours after any convenient time origin, the log-odds for ovulation are assumed to be In(p/q) = a + ¢~t;

(1)

it is further assumed that the effect of ultrasound is to produce the parallel line

ln(p]q) = a + 13(t + A),

(2)

with A representing a shift (reduction) in time due to an ultrasound effect. Estimated values (and standard errors) for the parameters a, /3, and h were obtained by standard m a x i m u m likelihood methods (see Anderson et al., 1980). Model fitting and assessment were done with the G L I M statistical software package (Baker and Nelder, 1978). Replicates were c o m b i n e d because there were no striking differences among them. III. R E S U L T S Ovulation was considered to have begun in a particular mouse if at least two ova were recovered. T h e f r e q u e n c y o f mice with only a single o v u m found, unilaterally, at autopsy was 2%. In our judgement, one-ovum recoveries can be artifacts due to mechanical rupture of a Graafian follicle by the dissecting instruments during oviduct removal. Evi-

488

Ultrasound in Medicine and Biology

dence drawn f r o m this study was the observation, in two mice, o f single ova (but never more) adhering to the outside o f an oviduct. T h e f r e q u e n c y o f oneo v u m mice was approximately equally divided between control and experimental groups. Classification of such mice as non-ovulators was a conservative approach that did not alter our interpretation of results. Data from Experiment 1 (adult mice, with h C G as the ovulatory h o r m o n e ) are presented in Table 2. At 12 and 13 h after h C G injection, ovulation had not yet begun in mice o f either the ultrasound- or sham-treated group. By 14 h after the h C G injection, ovulation was in progress at frequencies that did not differ significantly between the ultrasound and sham groups. Experiment l also revealed no demonstrable effect o f ultrasound treatment on n u m b e r o f ova observed, both at 14 and 15 h after hCG. Table 3 presents results o f Experiment 2 ( i m m a ture mice, e n d o g e n o u s L H as the o v u l a t o r y hormone). Ovulation was clearly in progress at 0400 h CT as seen from both the fractions of mice ovulating and the m e a n n u m b e r s of ova recovered. At none of the autopsy times was there a statistically significant difference between the ultrasound and sham treatm e n t groups either in fraction o f females having ovulated or in m e a n n u m b e r o f ova per female. Using the log,it model for analysis o f the frequency data in Experiment 2, we found a good fit o f the parallel-linelogit model. The estimated difference between exposed and sham groups in time o f ovulation (A) was 0.11 h _+ 0.33 h, with a 95% u p p e r confidence limit o f 0.65 h (and with/3 estimated as 1.18 _+ 0.25). Thus,

Table 2. Experiment 1 results: effect of ultrasound on timing of ovulation in adult strain BALB/c mice after follicle-priming injection of pregnant mare serum gonadotropin and ovulation-inducing injection of human chorionic gonadotropin (hCG) Timer of autopsy (no. hours after hCG)

Treatment

No. of Fraction mice of mice exposed ovulated

Mean no. ova per ovulated mouse (±SE)

12

Ultrasound Sham

7 7

0.00 0.00

---

13

Ultrasound Sham

10 10

0.00 0.00

---

14

Ultrasound Sham

10 10

0.10 0.30

(13) 6.0 ± 3.1

15

Ultrasound Sham

9 9

0.56 0.56

6.6 ± 2.7 4.6 ___1.2

t The times 14 and 15 h are represented in the text also as tt and &, respectively, for this experiment.

Volume 14, Number 6, 1988 Table 3. Experiment 2 results: effect of ultrasound on timing of ovulation in immature C129FI hybrid mice after follicle-priming injection of pregnant mare serum gonadotropin and release of endogenous ovulatory hormone (LH, under diurnal control) Timet of autopsy (Colony time)

Treatment

No. of Fraction mice of mice exposed ovulated

Mean no. ova per ovulated mouse (_SE)

0400

Ultrasound Sham

23 22

0.26 0.09

8.3 + 2.0 9.0 _+6.0

0500

Ultrasound Sham

22 22

0.50 0.55

15.1 ± 1.8 12.3 _ 1.7

0600

Ultrasound Sham None

22 22 17

0.68 0.73 0.71

14.3 ± 1.9 15.2 + 1.7 12.2 ___0.8

t The three autopsy times listed are represented in the text also as t~, t2, and t3 for this experiment.

for Experiment 2 we found no significant (p = 0.36) effect of ultrasound on time of ovulation. For a complete analysis, data on ovulation frequencies were c o m b i n e d from Experiments 1 and 2. D a t a c o m b i n i n g was w a r r a n t e d since for a given treatment Experiments 1 and 2 showed no significant differences in frequencies o f mice ovulating at the first two autopsies in which ovulation was seen; those autopsy times (t~ and &) are further described in the footnotes to Tables 1 and 2. (Although m e a n n u m b e r o f ova differed between the experiments, e.g. at autopsy time &, there was no basis on which to conclude that there was an associated difference in the time course, or rate, o f ovulation once it had begun.) The data on ovulation frequencies were c o m b i n e d as follows. For ultrasound-treated mice the t, autopsy results were added from Experiments 1 and 2, t2 autopsy results were similarly added, and t3 autopsy resuits were t a k e n f r o m E x p e r i m e n t 2. F o r s h a m treated mice the same data-combining m e t h o d was used. The statistical power o f the c o m b i n e d experiments m a y be estimated, based on the standard error o f A. There was a 97% (56%, respectively) probability of our finding a statistically significant result if in fact the true effect of ultrasound was to cause ovulation to occur as m u c h as one hour (one half hour, respectively) earlier than in sham control mice. (For comparison, in Experiment 2 the corresponding figures for statistical power were 92% and 45%, respectively.) The c o m b i n e d data, illustrated in Fig. 1, showed that the interval in time of ovulation between the ultrasound and sham groups was not statistically significant (A estimated as - 0 . 0 1 h ___0.28 h, with a 95%

Murine ovulatory response to ultrasound • A. H. GATES et al.

I

0 - SHAM. WITH FITTED LINE ( , , )

0

/

: 0

i 1 AUTOPSY TIME (HOURS AFTER FIRST 0VULATZON)

I

1

/

489

/

0

O' L 4'

-2

-3~

~

95~ CONFIDENCE LZMITS (---) FOR THE DIFFERENCE BETNEEN GROUPS

J 2

Fig. 1. Time course of ovul~ion in Experiments 1 and 2, combined: frequencies of mice having ovulated, in log odds (L) of ovul~ion, ~ con~cutive hourly inte~als. (Zero houB on the abcissa corresponds to t, in the text, ~r each experiment.)

upper confidence limit of 0.46 h and with/~ estimated as 1.19 _+ 0.23). IV. DISCUSSION A. Exposure level There are two principle mechanisms by which ultrasound is known to act on biological material: heating and cavitation. The temporal average intensities associated with almost all imaging techniques that might be used or obstetrics or gynecology are so small that it is reasonable to assume that heating of the tissues, and resultant effects, will be negligible. Acoustic cavitation, however, may occur and produce profound effects in lower organisms under the exposure conditions which are sometimes used in diagnosis (Child et al., 1981). The observed effects appear to be strongly dependent upon the temporal maximum intensity but not on the average intensity. Furthermore, there appear to be fundamental, qualitative differences between cavitation produced by continuous waves as opposed to very short pulses of ultrasound. If diagnostic ultrasound has any effect in human exposures, it possibly involves cavitation.

Our rationale in choosing exposure conditions for this project was twofold; first, we used temporal average intensities small enough that heating of the experimental animals would be negligible. (The maximum increase for our exposure conditions was 0.2°C in the region of the ovaries and 0.3°C directly ventral to the spine; temperatures were determined by a 50 #m copper-constantin thermocouple in an exposed mouse.) Second, we used pulse amplitudes and pulse lengths which would maximize the probability of cavitation occurring while approximating the conditions which might have relevance for clinical diagnosis. B. Ovulatory stage at exposure Testart et al. (1982) reported that premature ovulation had occurred when groups of patients were exposed at time intervals including 18-25 h after either hCG treatment or onset of the LH surge. At that exposure time, four out of six women ovulated within 36 h of gonadotropin stimulation--a timing considered by the authors to be abnormally early. Also, there was no difference between patients receiving one scan and those receiving two in frequency of such

490

Ultrasound in Medicineand Biology

early ovulations. Thus, one set of conditions those authors reported as being associated with premature ovulation in women was a single ultrasound exposure at, or shortly after, the midpoint between the time of occurrence of LH onset (or hCG injection) and the time of ovulation (as observed in controls). Consequently, to model our animal experiments after their human study, we exposed mice at 6-7 h after ovulatory hormone stimulation or about halfway between ovulatory hormone and the expected time of ovulation. The meiotic stage at the time we exposed mice was estimated as follows (based on Edwards and Gates, 1959 and Gates, 1969). At the earliest, some oocytes could have been in prometaphase-I (of the first meiotic division) if the retardation in onset of ovulation which we observed in sham groups (see Time o f ovulation, below) was due to delayed resumption of meiosis. At the latest, mice in Experiment 1 could have had oocytes in late metaphase-I if the average time of germinal vesicle breakdown was 1.5 h after hCG injection, as previously determined (Gates and Vadeboncoeur, 1970). All oocytes are assumed to have been exposed well before anaphase-I, which normally occurs about l0 h after hCG injection. In comparison, in the referenced study by Testart et al. (1982), those human exposures which were at 18-25 h after an hCG or LH stimulus possibly were near prometaphase-I; evidence is the persistence of the germinal vesicle for about 24 h (cited in Edwards and Steptoe, 1975).

C. Time of ovulation In mammals, the ovulation process, in which a mature ovum erupts from a Graafian follicle, is generally believed to be controlled by complex alterations in steroid metabolism, enzymatic pathways and inflammatory mediators (Yoshimura and Wallach, 1987). There is currently no reason to believe that there are fundamental differences in the mechanism of ovulation between mouse and man. Accordingly, it seems reasonable to expect that if preovulatory ultrasound were capable of causing premature ovulation in women, the same mechanism(s) should be operable in other mammals including the mouse. The immature mouse ovulation model (Experiment 2) proved more satisfactory for controlling time of ovulation than the adult, hCG-injected, mouse model (Experiment 1). The former procedure gave more between-run uniformity in percentage of mice ovulating and did not necessitate the inconvenient 0200 h hCG injection that was required in the adult animal model in order to schedule exposures and autopsies during convenient working hours.

Volume14, Number 6, 1988 In the present studies, ovulation occurred about 2 h later than expected. For example, in Experiment 2 ovulation began at 0400 h CT instead of at 0200 h as previously observed under similar conditions (Gates, 1969). In Experiment 1, ovulation also occurred about 2 h later than previously observed either in adult mice by Edwards and Gates (1959) or in prepubertal mice with hCG-induced ovulation (Gates, 1971). However, the time of ovulation among sham controls in the present experiment falls within the 2 h range of variability in time of endogenous ovulatory hormone release previously observed in C129F1 hybrids (Gates and Vadeboncoeur, 1970). Procedures having to do with sham or ultrasound treatment (e.g. pentobarbital treatment) were apparently not contributing factors in the unexpectedly late ovulation time in our studies since in Experiment 2, at the last autopsy time, the non-handled controls ovulated with essentially the same frequency as the sham-handled controls (see Table 3, last row). Our murine ovulatory tests were capable of detecting effects on time of ovulation, when either endogenous LH or injected hCG was used as the ovulatory hormone. The statistical analysis permitted detection of a difference between experimental groups in time of ovulation as great as one hour, with high statistical power. Specifically, the test used in Experiment 2 (with LH) and the tests used in Experiments 1 (with hCG) and 2, combined, had estimated probabilities of 92% and 97%, respectively, of detecting a one-hour shift in ovulation time. Thus, these experiments revealed no premature ovum release due to ultrasound exposure midway between ovulatory hormone stimulation and ovulation. The reported clinical findings which prompted this study (Testart et aL, 1982) raise questions about the accuracy with which differences in time of ovulation were assessible among women. For example, the time of ovulation which that study considered as being too early (i.e. less than 37 h after the ovulatory stimulus) was well within the limits for the normal time of ovulation after the LH rise, between 23.6 and 36.2 h (95% confidence interval), as estimated by other workers (WHO Task Force, 1980).

D. Number of ova Previous studies reported that average number of fetuses was reduced in rats exposed to ultrasound during proestrus and estrus (Bologne et al., 1983). The results were confirmed by Demoulin et al. (1985) who also investigated whether the litter size reduction could be due to a decrease in number of ova ovulated. The latter study subjected superovulating rats to ultrasound at approximately 1, 2, and 3 days before

Murine ovulatory response to ultrasound • A. H, GATES et al.

injection with ovulatory hormone (hCG). No reduction in ovum number occurred when exposure was before hCG treatment; however the question remained as to whether such reduction would result from ultrasound administered at a much later preovulatory stage, as is customary with gynecological monitoring to detect impending ovulation (e.g. for artificial insemination or in vitro fertilization procedures). The present experiments provide relevant information in that exposures were between the times of ovulatory hormone stimulus and ovulation. In our studies preovulatory exposure to ultrasound in the rodent did not reduce the number of ovulated ova. V. SUMMARY AND CONCLUSION Our investigations used mouse ovulation response tests (with both endogenous and exogenous sources of ovulatory hormone) to determine whether preovulatory ultrasound exposure, as used in gynecological applications, causes adverse temporal or quantitative effects on ovulation. Ultrasound and sham exposures were midway between the time of ovulatory hormone stimulation and the time of ovulation. Follicular oocytes were estimated to be between prometaphase and late metaphase of the first meiotic division at the time of exposure. The studies, both prospective and blind, provide no support for the postulates that, under controlled biological conditions relevant to human follicular growth monitoring, ultrasound treatment causes reduced numbers of ova ovulated or premature ovulation. Acknowledgments--This work was supported in part by NIH Grants HD 18932 and CA 39241, Program Project Grant ES 01248 and Center Grant ES 01247. We are grateful to Dr. Morton W. Miller for his helpful editorial suggestions. Also, we thank the following for their technical assistance: John Lockhart, Ellen Waiters, Rachel Berg, Catherine Crane, and Diane Dalecki.

REFERENCES Anderson S., Auquier A., Hauck W. W., Oakes D., Vandaele W. and Weisberg H. I. (1980) Logit analysis. In Statistical Methods for Comparative Studies (Edited by S. Anderson, A. Auquier, W. W. Hauck, D. Oakes, W. Vandaele, and H. I. Weisberg), pp. 161-167. John Wiley, New York. Baker R. J. and Nelder J. A. (1978) Generalized Linear Interactive Modeling--Release 3, Numerical Algorithms Group, Oxford. Bailey K. I., O'Brien W. D. Jr, and Dunn F. (1983) Ultrasonically induced morphological damage to mouse ovaries. Ultrasound in Med. & Biol. 9, 25-31. Bologne P., Demoulin A., Schaaps J. P., Hustin J. and Lambotte R. (1983) Influence des ultrasons sur la f6condit6 de la ratte. C. R. Soc. Biol. 177, 381-387. Bomsel-Helmreich O. (1976) The aging of gametes, heteroploidy, and embryonic death. Int. J. Gynaecol. Obstet. 14, 98-104. Bomsel-Helmreich O. (1985) Ultrasound and the preovulatory human follicle. Oxf. Rev. Reproduct. Biol. 7, 1-72. Campbell K. L. (1985) Methods of monitoring ovarian function

491

and predicting ovulation: Summary of a meeting. Res. Frontiers Fertil. Regulation 3(5), 1-16. Carstensen E. L. and Gates A. H. (1985) Ultrasound and the mammalian fetus. In Biological Effects of Ultrasound (Edited by W. L. Nyborg and M. C. Ziskin), pp. 85-95. Churchill Livingstone, New York. Child S. Z., Carstensen E. L. and Lam S. K. (1981) Effects of ultrasound on Drosophila: III. Exposure of larvae to low-temporal-average-intensity, pulsed irradiation. Ultrasound in Med. & Biol. 7, 167-173. Demoulin A. (1985) Is diagnostic ultrasound safe during periovulatory period? Res. in Reprod. 17(2), 1-2. Demoulin A., Bologne J. and Lambotte R. (1985) ls ultrasound monitoring of follicular growth harmless? Ann. N Y Acad. Sci. 442, 146-152. Edwards R. E. and Gates A. H. (1959) Timing of the stages of the maturation divisions, ovulation, fertilization and the first cleavage of eggs of adult mice treated with gonadotrophins. J. Endocrinol. 18, 292-304. Edwards R. E, and Steptoe P. (1975) Induction of follicular growth, ovulation and luteinization in the human ovary. J. Reprod. Fert. (Suppl.), 121-163. Gates A. H. (1969) Time relationship between neural stimulus for LH-releasing factor and ovulation in the PMS-treated mouse. Am. Zoologist 9, 1080. Gates A. H. (1971) Maximizing yield and developmental uniformity of eggs. In Methods in Mammalian Embryology (Edited by J. C. Daniel, Jr.), pp. 64-75. W. H. Freeman & Co., San Francisco. Gates A. H. and Bozarth J. L. (1978) Ovulation in the PMSGtreated immature mouse: effect of dose, age, weight, puberty, season and strain (BALB/c, 129 and C I29Ft hybrid). Biol. Reprod. 18, 497-507. Gates A. H. and Vadeboncoeur M. E. (1970) Timing of meiosis and ovulation following endogenous or exogenous LH stimulation in the PMS-treated prepuberal mouse. Third Annual Meeting of Soc. for Study of Reproduction, Columbus, Ohio. Abstracts of Papers, p. 19. Academic Press, New York. Hansmann M., Hackeloer B.-J., Staudach A. and Feichtinger W. (1986) Sonographic aspects of the menstrual cycle. In Ultrasound Diagnosis in Obstetrics and Gynecology (Edited by M. Hansmann, B.-J, Hackeloer, and A. Staudach), pp. 361-382. Springer-Verlag, Berlin. Marinho A. O., Sallam H. N., Goessens L. K. V., Collins W. P., Rodeck C. H. and Campbell S. (1982) Real time pelvic ultrasonography during the periovulatory period of patients attending an artifical insemination clinic. Fertil. Steril. 37, 633-638. O'Brien W. D. Jr. (1985) Biological effects in laboratory animals. In Biological Effects of Ultrasound (Edited by W. L. Nyborg and M. C. Ziskin), pp. 77-84. Churchill Livingstone, New York. Puissant F., Lejeune B. and Leroy F. (1984) Effects on mature mouse oocytes of ultrasound treatment applied before in vitro fertilisation. Intern. Res. Commun. Serv. (Med. Sci.) 12, 421-422. Stewart H. F. (1983) Output levels from commercial diagnostic ultrasound equipment. J. Ultrasound Med. (Suppl.) 2(10), 39. Test,art J., Thebault A., Souderes E. and Frydman R. (1982) Premature ovulation after ovarian ultrasonography. Br. J. Obstet. Gynaecol. 89, 694-700. Thacker J. (1985) Investigations into genetic and inherited changes produced by ultrasound. In Biological Effects of Ultrasound (Edited by W. L. Nyborg and M. C. Ziskin), pp. 67-76. Churchill Livingstone, New York. WHO Task Force. (1980) Temporal relationships between ovulation and defined changes in the concentration of plasma estradiol-17/~, luteinizing hormone, follicle-stimulating hormone, and progesterone: I. Probit analysis. Am. J. Obstet. Gynecol. 138, 383-390. Yoshimura Y. and Wallach E. E, (1987) Studies of the mechanism(s) of mammalian ovulation. Fertil. Steril. 47, 22-34. Ziskin M. C. (1986) The American Institute of Ultrasound policy and statement on safety. Ultrasound in Med. & Biol. 12, 711-714.