FERTILITY AND STERILITY Copyright 1993 The American Fertility Society Printed on acid-free paper in U. S. A. Ovulation after intravenous and intramuscular human chorionic gonadotropin*t Robin A. Fischer, M.D.:\: Steven T. Nakajima, M.D. Mark Gibson, M.D. II John R. Brumsted, M.D. Division of Reproductive Endocrinology and Infertility, Department of Obstetrics and Gynecology, University of Vermont College of Medicine, Burlington, Vermont Objective: To define the time interval from intravenous and intramuscular hcg administration to follicular wall rupture and the endocrinologic events associated with ovulation. Design: Subjects were studied in two cycles and received hcg either 10,000 IU 1M or 500 IU IV in a random sequence with an intervening spontaneous menstrual cycle. Patients: Thirty women from the University of Vermont Reproductive Endocrinology Service with unexplained, male, or cervical factor infertility. Interventions: Subjects underwent superovulation with clomiphene citrate followed by hcg administration when the lead follicle reached a mean diameter of 18 mm. Follicular rupture was determined by ultrasound monitoring every 2 hours starting 31 and 30 hours after intravenous and intramuscular hcg, respectively. Serum samples were obtained hourly for hormone measurements. The study was completed 2 hours after follicular rupture or 48 hours after hcg administration. Results: Twenty-five subjects received both intramuscular and intravenous hcg. The mean time to ovulation was 40.4 hours after intramuscular hcg (range,,,;36 to 48 hours) and 38.3 hours after intravenous hcg (range, 33 to 48 hours). No differences were noted in the time interval to ovulation or rate of change in circulating E2 and P levels after 1M versus IV hcg administration. Conclusions: These findings suggest (1) ovulation occurs over a broad range of time after hcg administration; (2) ovulation does not occur in a more specific time interval after intravenous than intramuscular hcg; and (3) the rate of change in circulating E2 and P levels are not different after intravenous than intramuscular hcg. Fertil SteriI1993;60:418-22 Key Words: HCG, ovulation, IV, 1M, superovulation In spontaneous and controlled ovarian hyperstimulation cycles, ovulation occurs 36 to 38 hours Received December 10, 1992; revised and accepted May 25, 1993. * Supported in part by the University of Vermont General Clinical Research Center, National Institutes of Health, Division of Research Resources (GCRC RR109), Burlington, Vermont. t Presented in part at the 47th Annual Meeting of The American Fertility Society, Orlando, Florida, October 21 to 24, 1991. :j: Reprint requests and present address: Robin A. Fischer, M.D., Division of Reproductive Endocrinology and Infertility, Department of Obstetrics and Gynecology, University of Massachusetts Medical Center, 55 Lake Avenue North, Worcester, Massachusetts 01655. 418 Fischer et al. Ovulation after IV and 1M HCG after the onset of the LH peak (1-3). From a prior study (4), the onset of the spontaneous LH surge is abrupt, with LH levels doubling within 2 hours, and peak levels are sustained for a period of 12 to 14 hours. Because the exact onset of LH peak is difficult to determine, the onset of the LH surge appears to be a more reliable reference point (1, 2, 4). Human chorionic gonadotropin is similar in Present address: Division of Reproductive Biology and Medicine, University of California at Davis, Davis, California. II Present address: Department of Obstetrics and Gynecology, Health Sciences Center, West Virginia University, Morgantown, West Virginia.
structure to LH and has been used to induce ovulation in most of the assisted reproductive procedures. The time interval between an intramuscular hcg injection and the physical and endocrinologic events of ovulation have not been defined. It has been presumed that intramuscular hcg administration gives serum levels similar to the onset of an LH surge, and' therefore ovulation occurs approximately 36 hours later. A previous study has reported that intramuscular hcg injection leads to variable serum levels because of erratic absorption (Nader S, Berkowitz AS, Held B, Winkel CA, abstract). Therefore, one might expect a large variation in the time interval from intramuscular hcg administration to ovulation. The purpose of this study was to define the range of time over which the biochemical and physical events of ovulation occur after intramuscular or intravenous hcg injection. Human chorionic gonadotropin administration by the intravenous route results in very precise and immediate serum hcg levels (Nakajima ST, Stewart DR, Overstreet JW, Lasley BL, abstract). This study was initiated to define the time interval from hcg administration to ovulation and to test the hypothesis that more predictable serum levels of hcg achieved by an intravenous dose results in decreased variability in the time interval between hcg administration and ovulation. Further, the pattern of change in serum E2 and P concentrations, relative to hcg injection and follicular rupture, were described. MATERIALS AND METHODS Thirty women volunteers, recruited from the University of Vermont Reproductive Endocrinology and Infertility clinic with unexplained, male, or cervical factor infertility were studied in the General Clinical Research Center after informed consent was obtained. Subjects were already receiving or planning to undergo superovulation with clomiphene citrate (CC), followed by hcg administration, with timed washed lui, as a therapy for their infertility. A standard course of CC (50 mg orally cycle days 5 to 9) was administered, followed by an hcg injection when the lead ovarian follicle reached a diameter of 18 mm, as determined by transvaginal ultrasound (US) (diameter is the mean of the 3 largest diameters, each measured in a different plane). Subjects were studied in two cycles and received hcg either by intramuscular injection (10,000 IU) or intravenous injection (500 IU) in a random sequence with an intervening menstrual cycle. Cycles began with a baseline transvaginal US performed between cycle day 1 and 5, to ensure ovarian inactivity. Daily blood samples and twice daily (morning and early evening) urine LH testing (Monoclonal Antibodies, Mountain View, CA and Quidel, San Diego, CA) began on cycle day 10, to ensure the absence of a spont;meous LH surge. A second transvaginal US was performed on cycle day 10 to measure ovarian follicular growth. From this US, the day of expected hcg administration was predicted by extrapolating follicular growth as 2 mm/d (5). Subjects returned on the day of expected hcg administration (at approximately 8:00 A.M.) for confirmation of follicular size. Immediately after US documentation of at least one ovarian follicle with a mean diameter of 18 mm, the daily blood sample was obtained and hcg was administered. The first 18 subjects were admitted to the General Clinical Research Center 36 and 21 hours after intramuscular and intravenous hcg administration, respectively. All subsequent subjects were admitted to the General Clinical Research Center 30 and 31 hours after intramuscular and intravenous hcg administration, respectively. Initially, US monitoring began 36 hours after intramuscular hcg but was changed to begin 30 hours after intramuscular hcg after five subjects had evidence of follicular rupture before 36 hours after intramuscular hcg. Transvaginal US images were obtained using an ADR real-time scanner and a 7.5-mHz endovaginal probe (ATL/ADR Ultrasound, Bothel, W A). Ultrasound examinations were performed every 2 hours seeking evidence of follicular rupture (ovulation). The study was completed 2 hours after follicular rupture was documented or 48 hours after hcg administration, at which time a washed lui was performed. Follicular rupture was defined by either of the following two criteria: [1] the disappearance of the follicle with or without accumulation of fluid in the cul-de-sac or [2] change of the follicle from clear with regular borders to echogenic and noticeably irregular borders (6, 7). Serum samples were obtained every 2 hours for the first 20 subjects and increased in frequency to every hour for the last 10 subjects. Samples were collected from an indwelling IV catheter for determination of E2 and P levels. Subjects were kept at bed rest with bathroom privileges. Meals were ingested immediately before venous sampling, to minimize alterations in clearance of steroids, affecting serum steroid levels (8). Fischer et al. Ovulation after IV and 1M HCG 419
All blood samples were centrifuged within 2 hours of collection and stored at -20 C until assayed. Steroid concentrations in all blood samples from a given subject were determined within assay. Estradiol and P plasma concentrations were determined by RIA (Leeco Diagnostics, Inc., Southfield, MI). Intra-assay and interassay coefficients ofvariations (CVs) were 8.8% and 7.5% for E2 and 3.7% and 7.6% for P, respectively. Plasma LH and hcg levels were determined by time-resolved immunofluorimetric sandwich kit assays (DELFIA; Pharmacia Diagnostics, Inc., Columbia, MD). Intra-assay and interassay CVs were 5.1% and 12.8% for LH and 4.3% and 5.8 for hcg, respectively. The hcg assay is calibrated against the WHO First International Reference P,eparation of hcg (WHO 1st IRP 75/537). The time interval from hcg administration to follicular rupture after intramuscular and intravenous hcg were compared using Wilcoxon's signed rank test. Alterations in E2 and P serum concentrations were evaluated with Wilcoxon's signed rank test. 17-(3 E2 concentrations were compared using a percent change ratio: (E2 concentration at a specific time - E2 concentration at the time of hcg)/e2 concentration at the time ofhcg. This was done to exclude any influence of multiple follicle development in a studied cycle on E2 concentration. Progesterone concentrations were compared directly because some P levels were below the limit of assay detection, and therefore a calculated ratio comparison was impossible. Periovulatory serum P concentrations were compared using analysis of variance. The administered intramuscular and intravenous hcg doses were chosen using the following rationale. A 10,000 IU dose ofhcg is widely used in clinical practice to induce ovulation. A 500 IU IV hcg dose was determined based on an hcg volume of distribution = 3 liters (Gibson M, Nakajima S, unpublished data). An IV dose ofhcg (in IU), three times the usual peak LH concentration (50 to 200 IU/L), should simulate the LH peak. RESULTS z w fi) ID 0 Thirty subjects were enrolled. Three subjects conceived in their first study cycle and therefore did not have a second study cycle for comparison. Two subjects conceived spontaneously during their intervening cycle and therefore had data available for only one study cycle. Twenty-five subjects finished both study cycles, and 15 had a specific time of ovulation documented by vaginal US in two study cy- I&- 0 ::l..j U U 30 <36 36 38 38-40 40-42 42-44 44-46 46-48 TOTAL INTERVAL FROM hcg ADMINISTRATION TO OVULATION (HOURS) Figure 1 Frequency distribution of the time interval from hcg administration to follicular wall rupture after intramuscu lar hcg (0, n = 23) and intravenous hcg (III, n = 25) in 2-hour intervals. cles. In 5 subjects, specific time of follicular rupture was not noted because ovulation occurred before the first planned vaginal US after intramuscular hcg. None ofthese 5 subjects demonstrated a spontaneous LH surge before hcg administration during the study cycle by urine LH ovulation predictor kit; this was later confirmed by serum LH ::::;14 IU/L in all cycles. Six subsequent subjects failed to show ultrasonographic evidence of follicular rupture, up to 48 hours after hcg administration (3 after intramuscular hcg, 2 after intravenous hcg, and 1 after both intravenous and intramuscular hcg). Thus, of the 30 subjects studied in 55 cycles, US documentation of follicular rupture occurred during the study interval in 78.0% (43/55) of cycles. If one includes the cycles in which US documented follicular rupture before the first monitoring exam, evidence of follicular rupture was noted by US in 87.3% (48/55) of cycles (Fig. 1). The mean time to ovulation after intramuscular and intravenoul'l hcg in women with documented ovulation during both study cycles is 39.8 hours and 37.4 hours, respectively (data used for Wilcoxon's signed rank test). The mean time to ovulation after intramuscular and intravenous hcg in all study cycles with US documented evidence of follicular rupture was 40.4 and 38.3 hours, respectively. Median time to follicular rupture for all subjects was calculated to be 38 and 37 hours after intramuscular and intravenous hcg administration, respectively. The difference between the mean time to follicular rupture after intramuscular and intravenous hcg was not statistically different (0.05 P::::; 0.10, Wilcoxon's signed rank). Frequency analysis of hours to ovulation after intramuscular hcg demonstrated that 30% of 420 Fischer et at. Ovulation after IV and 1M ReG
subjects ovulated by 36 hours, 59% by 40 hours, and 85% by 46 hours after intramuscular hcg. The time to ovulation after intramuscular hcg ranged from <36 hours to >48 hours. Frequency analysis of hours to ovulation after intravenous hcg shows 20% of subjects ovulated by 36 hours, 70% by 40 hours, and 87% by 46 hours after intravenous hcg. The time to ovulation after intravenous hcg ranged from 33 hours to >48 hours. Human chorionic gonadotropin levels were measured in four randomly selected subjects. Before either intramuscular or intravenous hcg administration, hcg levels were <2 lull. Thirty-six hours after intramuscular hcg, serum hcg levels were 201 to 359 lull (mean, 311 lull). Twenty-one hours after intravenous hcg, serum hcg levels were 36 to 55 lull (mean, 45 lull); 35 hours after intravenous hcg, serum hcg levels were 17 to 28 lull (mean, 23 lull). Patterns of E2 and P concentrations in the periovulatory period were analyzed in the subset of subjects with evidence of follicular rupture during the study interval in both study cycles (n = 15, 30 cycles). Serum E2 and P concentrations for each subject were standardized to the time of ovulation documented by vaginal US (time zero = ovulation). The rate of change in E2 and P levels from the time of hcg administration until documented ovulation was not statistically different in the intramuscular versus the intravenous hcg cycles (P 0.10, Wilcoxon's signed rank). Thirty hours after hcg administration, E2 levels were noted to decrease, whereas P levels increased. 17 -{3 E2 levels decreased from the time of intramuscular and intravenous hcg administration until the time of ovulation by a mean 63% and 61 %, respectively (P 0.05, paired t-test). The range of decrease in E2 varied from 78% to 28% after intramuscular hcg and 75% to 37% after intravenous hcg. Progesterone levels increased from the time of intramuscular and intravenous hcg administration until the time of ovulation by a mean of 2.4 nmoljl and 2.6 nmoljl, respectively. Maximum and minimum changes in P after intramuscular hcg were an increase of 5.6 nmoljl and a decrease of 2.0 nmol/l, respectively. Maximum and minimum changes in P after intravenous hcg were an increase of 9.3 nmoljl and a decrease of 0.4 nmoljl, respectively. There was a sharp and statistically significant (P 0.05, ANOV A) increase in serum P concentration immediately after follicular rupture (Fig. 2). 1.4 +---------------------t 1.2 0.2-4 -2 +2 TIME RELATIVE TO OVULATION (HOURS) Figure 2 Mean (±SE) concentrations of serum P in the periovulatory period. Time zero is time of ultrasonographic evidence of follicular rupture. (1.0 ng P ImL = 3.18 nmol P IL). DISCUSSION Ovulation occurs over a wide range of time after both intramuscular and intravenous hcg administration. The time to ovulation is <36 hours to >48 hours after intramuscular hcg and 33 hours to >48 hours after intravenous hcg. Recognition of the variability in the interval to ovulation after hcg administration may explain why reported pregnancy rates with superovulation/hcg/lui are not as high as with other assisted reproductive technologies. Statistically, the time to ovulation does not appear to be more predictable after intravenous rather than intramuscular hcg administration and thus rejects our original hypothesis. The observation of a broad time interval from hcg administration to follicular wall rupture is unique. The reason for such variance is yet to be determined. These data do not support variability in serum hcg levels after intramuscular hcg administration as a primary cause for the observed variance. Follicles may vary in their sensitivity to hcg because of differences in LH receptor content or a discordance between follicular size and hcg sensitivity. Furthermore, the effect ofhcg may not be readily predicted by measurable criteria like follicle size or hormone levels but predicted only by follicular microenvironment parameters such as cellular LH receptor concentration andlor sensitivity. Further investigation is necessary to fully explain our observations. Four of the 30 subjects conceived during a study cycle. One of these subjects failed to demonstrate evidence of follicular rupture after intramuscular hcg but did ovulate 39 hours after IV hcg and conceived during that cycle. Two others conceived during their first study cycle with evidence of follic- Fischer et al. Ovulation after IV and 1M HCG 421
ular rupture 40 and 46 hours after intramuscular hcg. The final subject did conceive after intravenous hcg administration but failed to demonstrate ultrasonographic evidence of follicular rupture. An early obstetrical US, however, was consistent with conception 1 week after the monitored study period. These observations may support the need for increased ultrasonographic surveillance in some patients failing initial treatment with this therapy. Unfortunately, ultrasonographic monitoring did not extend beyond 48 hours after hcg administration. Lack of evidence for ovulation occurred in 6/55 cycles (>10%). This apparent absence ofovulation may be related to the following: [1] an extended ovarian response interval to hcg administration; [2] the inability to distinguish between a corpus luteum (CL) and a follicle by US; [3] spontaneous occurrence of the luteinized unruptured follicle syndrome; or [4] premature administration of hcg to immature unresponsive follicles. Alterations in the rate of change in circulating E2 and P levels were not different after intramuscular versus intravenous hcg. 17 -f3 E 2 1evels rose steadily before and quickly decreased after hcg administration. Progesterone levels appear to remain low and unchanged until hcg administration, and then began to steadily increase. These sex steroid changes after intramuscular or intravenous hcg administration closely approximate those observed after an endogenous LH surge in an unstimulated cycle (4). The modest rise in serum P concentration after follicular rupture is another unique observation. This rapid increase in serum P level may be due to either release from a follicular fluid (FF) inhibiting factor or rapid transperitoneal absorption of P present in FF (9, 10). It is unlikely that the rapid increase in serum P concentration is from the newly formed CL, which should not be well vascularized immediately after follicular wall rupture. Collectively, these data serve to define the wide range of time to ovulation after hcg administration. Ovulation does not occur in a more specific or predictable time interval after intravenous than intramuscular hcg injection. This finding refutes variable serum hcg levels as the cause for the wide time range from hcg dose until ovulation (Nader S, Berkowitz AS, Held B, Winkel CA, abstract). The lack of normal fertile controls limits the conclusions that can be drawn from these data. Because we studied an infertile group, they may have had additional unidentified periovulatory abnormalities. Further, these data define the pattern of E2 and P production in response to hcg-stimulated ovulation. These observations confirm that hcg administration results in stroid changes similar to those noted after a physiological LH surge. Acknowledgments. We thank the following from the University of Vermont, Burlington, Vermont: Susan Lewis, A.S., M.L.T., for her technical assistance; Gary J. Badger, M.S., for statistical assistance; and Ms. Susan Lapworth for her assistance in preparation of this manuscript. We appreciate the generous material support of ADR/ATL, Bothel, Washington; Quidel, San Diego, California; and Monoclonal Antibodies, Mountain View, California. REFERENCES 1. Taymor ML, Seibel MM, Smith D, Levesque L. Ovulation timing by luteinizing hormone assay and follicular puncture. Obstet GynecoI1983;62:191-5. 2. Testart J, Frydman R. Minimal time lapse between luteinizing hormone surge or human chorionic gonadotropin administration and follicular rupture. Fertil SteriI1982;37:50-3. 3. Yussman MA, Taymor ML. Serum levels offollicular stimulating hormone and luteinizing hormone and of plasma progesterone related to ovulation by corpus luteum biopsy. J Clin Endocrinol Metab 1970;30:396-9. 4. Hoff JD, Quigley ME, Yen SSC. Hormonal dynamics at midcycle: a reevaluation. J Clin Endocrinol Metab 1983;57:792-6. 5. Hackeltier BJ, Fleming R, Robinson HP, Adam AH, Coutts JRT. Correlation of ultrasonic and endocrinologic assessment of human follicular development. Am J Obstet Gynecol 1979;135:122-8. 6. Sanders RC, James AE Jr. The principles and practice of ultrasonography in obstetrics and gynecology. 3rd ed. Norwalk (CT): Appleton-Century-Crofts, 1985:523. 7. Callen PW. Ultrasonography in obstetrics and gynecology. 3rd ed. Philadelphia (PA): W.B. Saunders, Co., 1988:428. 8. Nakajima ST, Gibson M. The effect of a meal on circulating steady-state progesterone levels. J Clin Endocrinol Metab 1989;69:917-9. 9. Gibson M, Samach A, Brumsted JR, Auletta FJ. Fate of peritoneal progesterone in the rabbit. Steroids 1985;46:735-40. 10. Koninckx PR, Heyns W, Verhoeven G, VanBaelen H, Lissen WD, DeMon P, et al. Biochemical characterization of peritoneal fluid in women during the menstrual cycle. J Clin Endocrinol Metab 1980;51:1239-44. 422 Fischer et al. Ovulation after IV and 1M HCG