Department of Obstetrics and Gynecology, The University of Texas Health Science Center at San Antonio, San Antonio, Texas

Transcription

1 FERTILITY AND STERILITY Copyright <> 1991 The American Fertility Society Vol. 56, No. 2, August 1991 Printed on ocid-free paper in U.S.A. Follicular size at the time of human chorionic gonadotropin administration predicts ovulation outcome in humaij menopausal gonadotropin-stimulated cycles* Kay len M. Silverberg, M.D. t David L. Olive, M.D. William N. Burns, M.D. Julia V, Johnson, M.D.:j: Terry R. Groff, M.D. Robert S. Schenken, M.D. Department of Obstetrics and Gynecology, The University of Texas Health Science Center at San Antonio, San Antonio, Texas Objective: The objectives of this study were: (1) to correlate follicle size by transvaginal sonography with ovulation outcome in cycles of controlled ovarian hyperstimulation with human menopausal gonadotropins; (2) to determine iffollicular size on the day of human chorionic gonadotropin (hcg) administration predicts the incidence of ovulation; and, if so, (3) to derive a mathematical model that predicts the number of expected ovulations in any given cycle of controlled ovarian hyperstimulation. Design: A retrospective analysis. Participants: Forty-nine consecutive patients undergoing 122 cycles of controlled ovarian hyperstimulation were studied in a tertiary care setting. Main Outcome Measures: Follicular size and evidence of ovulation were determined sonographically. The main outcome measure was the rate of ovulation per follicle size. Results: The percentages of follicles measuring :5:14 mm, 15 to 16 mm, 17 to 18 mm, 19 to 20 mm, and >20 mm on the day ofhcg administration that subsequently ovulated were 0.5%, 37.4%, 72.5%, 81.2%, and 95.5%, respectively. Conclusions: (1) Follicular size on the day of hcg administration correlates with the incidence of ovulation. (2) The expected number of ovulations in any given controlled ovarian hyperstimulation cycle can be predicted with 95% confidence using the accompanying equation. Fertil Steril 56:296, 1991 The criteria for timing human chorionic gonadotropin (hcg) administration during cycles in which controlled ovarian hyperstimulation is achieved with human menopausal gonadotropin (hmg) are poorly defined. Several guidelines have been suggested, including serum estradiol (E 2) levels of 200 to 300 pg/ml per mature follicle, and mean follicular diameter at least 14 mm as determined by ultrasonog- Received January 2, 1991; revised and accepted April 3, * Presented at the 46th Annual Meeting of The American Fertility Society, Washington, D.C., October 13 to 18, t Reprint requests: Kaylen M. Silverberg, M.D., Division of Reproductive Endocrinology and Infertility, Department of Obstetrics and Gynecology, The University of Texas Health Science Center at San Antonio, 7703 Floyd Curl Drive, San Antonio, Texas :j: Present address: Department of Obstetrics and Gynecology, University of Vermont Medical School, Burlington, Vermont. 296 Silverberg et al. FoUicular size predicts ovulation raphy. 1-5 Today, most clinicians employ a combination of serum E 2 level and ultrasonographic follicular size criteria. 6 The ultrasonographic criteria for timing hcg administration have been suggested to assure the release of mature oocytes capable of being fertilized and developing into a viable pregnancy and to limit the occurrence of multiple pregnancy. These guidelines rely on the assumptions that oocyte maturity correlates with follicular size and that all follicles that attain a certain size will ovulate. Although there are some data from in vitro fertilization -embryo transfer (IVF-ET) programs correlating follicle size with oocyte maturity, 7 there is no information correlating follicle size with ovulation during controlled ovarian hyperstimulation. We therefore correlated follicle size on serial transvaginal sonograms with ovulation outcome in

2 patients undergoing controlled ovarian hyperstimulation with hmg. Our goals were to assess if follicular size at the time of hcg administration predicts the incidence of ovulation and, if so, to derive a mathematical model that would allow prediction of the total number of ovulations in any given cycle of controlled ovarian hyperstimulation. 1 MATERIALS AND METHODS Forty-nine consecutive patients undergoing 122 cycles of controlled ovarian hyperstimulation at the Center for Reproductive Medicine at Humana Women's Hospital-South Texas were studied. Patient diagnoses included ovulatory dysfunction (n = 21), unexplained infertility (n = 11), endometriosis (n = 8), male factor (n = 2), and two or more infertility factors (n = 7). Two days after the onset of menses, patients underwent transvaginal sonography using an ATL Ultramark 4 equipped with a 5-MHz endovaginal probe (Advanced Technology Laboratories, Bothel, W A). Ovarian stimulation was initiated using hmg (Pergonal; Serono Laboratories, Inc., Randolph, MA) if the largest follicle was <10 mm. All cycles were initiated using 150 international units (IU) of hmg daily for 3 days. On cycle day 6, patients were again seen for a transvaginal sonogram and a serum E 2 level. Thereafter, patients were seen every other day between 8 and 9 A.M. for transvaginal sonograms and serum E 2 levels. When the leading follicle(s) achieved a mean diameter of 13 mm, all patients were seen on a daily basis. The hmg dose was not altered as long as serial E 2 levels rose between 50% and 100%. The dosage was reduced by 75 IU if two consecutive E 2 levels rose by > 100%. Human chorionic gonadotropin (Profasi; Serono Laboratories, Inc.) was administered in a single intramuscular dose of 10,000 IU to trigger ovulation. Intercourse or intrauterine insemination (lui) was performed 34 to 35 hours after hcg administration. Intrauterine insemination was performed only for unexplained infertility and for cervical factor infertility. Patients with male-factor infertility were treated with either cervical or lui using cryopreserved donor sperm. Daily ultrasound (US) examination was performed until at least 42 hours after hcg administration. None of the patients underwent oocyte retrieval. Definitive ultrasonographic evidence of ovulation was defined by either disappearance or obvious collapse of a follicle or formation of an echogenic corpus luteum. Because the US examinations were performed by all of the authors, we developed a test to evaluate interexaminer mea- surement variation. Twenty-three follicles varying in size from 8 to 21 mm were videotaped, and each investigator measured each follicle. The overall interexaminer coefficient of variation ( CV) was 5.1%. We therefore grouped follicles in 2 mm increments for data analysis to correct a possible 5% (1 mm) error. No luteal phase hormonal support was administered, and a qualitative serum fj-hcg was performed no earlier than 14 days after ovulation for those women who had not started their menses. Transvaginal sonography was performed 4! weeks after ovulation to document fetal viability in all patients having a positive serum fj-hcg test. Patients who did not conceive underwent transvaginal sonography 2 days after the onset of subsequent menses to rule out the presence of residual follicular cysts. Residual cysts were defined as sonolucent follicles measuring >15 mm. Preclinical pregnancy was defined as a temporal increase in serum fj-hcg level, without ultrasonographic evidence of an intrauterine pregnancy (IUP). Clinical pregnancy was defined as either an ectopic pregnancy (EP), an IUP with cardiac activity documented with transvaginal sonography, or chorionic villi identified in a pathology specimen. Ongoing/ delivered pregnancy includes ongoing clinical pregnancies with documented cardiac activity and pregnancies that have resulted in delivery of viable infants. Daily serum samples were assayed for E 2 using commercially available radioimmunoassay kits (Diagnostic Products Corporation, Los Angeles, CA) according to the manufacturer's directions. Interassay and intra-assay CVs were 6.4% and 2.7%, respectively. The assay sensitivity was 5 pg/ml. Data were analyzed using Student's t-test for two sample interval comparisons for normally distributed data. Pearson's r was used for correlations, and multiple linear regression analysis was used to determine the relative contribution of any given follicle to total serum E 2 level, based on its size on the day of hcg administration. RESULTS Ultrasound confirmed both the presence of ovulation and the number of follicles that ovulated in 103 of122 cycles (84.4%). We were unable to confirm ovulation in 1 cycle, and we were unable to determine the exact number of ovulations in 18 cycles. A total of 1,344 follicles were studied. For this analysis, we included any follicle measuring 6 mm or larger on the day of hcg administration in this study. Vol. 56, No. 2, August 1991 Silverberg et al. Follicular size predicts ovulation 297

3 Table 1 Incidence of Ovulation by Follicular Size on the Day of HCG Administration No. of No. of follicles ovulations Ovulations Follicule size ::;;;14mm to 16 mm to 18 mm to 20 mm >20mm Total 1, Follicles ~ 17 mm ovulated significantly more often than did those ::;;;16 mm (P < 10-7 ). Clinical pregnancies occurred in 19 of 103 cycles (18.4%). Twelve pregnancies (63.2%) are currently ongoing or have already resulted in the delivery of viable infants. Of these, 6 were singleton gestations, and 6 were multiple gestations. Six additional pregnancies (31.6%) resulted in spontaneous miscarriage, and 1 (5.2%) was an EP. Overall, 243 of the 1,344 follicles ovulated (18.1%) (Table 1). Follicles measuring 17 mm or larger on the day of hcg administration ovulated significantly more often than did those follicles measuring 16 mm or smaller on the day of hcg administration (P < 10-7 ). Follicles measuring 14 mm or smaller rarely ovulated. There was no significant difference in the incidence of ovulation among those follicles measuring 17 to 18 mm, 19 to 20 mm, and 21 mm or larger on the day of hcg administration. An equation was derived to predict the number of follicles that will ovulate in a given cycle based on their size on the day of hcg administration. The expected number of ovulations equals 0.96 (A) (B) (C) (D) (E). The 95% confidence interval for this equation was constructed. The upper limit of this interval for expected number of ovulations was: 1.00 (A) (B) (C) (D) (E), and the lower limit was: 0.87 (A) (B) (C) (D) (E). The letters A, B, C, D, and E refer to mean follicular diameters of >20 mm, 19 to 20 mm, 17 to 18 mm, 15 to 16 mm, and s;14 mm, respectively. Using multiple linear regression analysis, a second equation was derived to determine the relative contribution of any given follicle to total serum E 2 level, based on its size on the day of hcg administration. The serum E 2 level equals 291 pg/ml (X) + 64 (Y) (Z). The letters X, Y, and Z refer to mean follicular diameters of ~17 mm, 15 to 16 mm, and s;14 mm, respectively. No significant dif- % ference in E 2 production was detectable for follicles measuring 14 mm or smaller. Likewise, when a follicle reached 17 mm, its contribution to total serum E 2 level remained relatively constant, irrespective of further growth. Although serum E 2 levels correlated with the total number of follicles present on the day of hcg administration (r = 0.433, P < 0.01), there was no correlation between serum E 2 level and the number of follicles that ovulated. The mean serum E 2 level for all cycles on the day of hcg administration was 947 pg/ml. No significant difference was noted between the mean serum E 2 level in cycles resulting in ongoing pregnancy (958 pg/ml), spontaneous miscarriage (1,043 pg/ml), or failed conception (953 pg/ml). There was no correlation between the maximum dose of hmg administered and the number of ovulatory follicles. In addition, there was no relationship between the total hmg dose administered and either the total number of follicles present on the day of hcg administration or the number of follicles that subsequently ovulated~ The number of ovulations did not correlate with fecundity (Table 2). The total number of follicles that developed and the total number of ovulations that occurred were not significantly different between conception and nonconception cycles (Table 3). There was no significant difference in the number of large follicles (> 15 mm) on the day of hcg administration in cycles resulting in conception and those not resulting in conception. The chance of a specific-size follicle ovulating was the same for conception and nonconception cycles (Table 4). There was no relationship between the size of the largest ovulating follicle and cycle outcome; patients whose largest ovulatory follicle was 16 mm did not have a significantly lower incidence of conception when compared with patients whose largest ovulatory follicle was > 16 mm. In addition, the occurrence of Table 2 Incidence of Conception by Number of Ovulations No. of patients No. of pregnancies No. of ovulations 1 28 (27.2)" 3 (10.7)b 2 32 (31.1) 8 (25.6) 3 24 (23.3) 5 (20.8) 4 15 (14.5) 2 (13.3) 5 3 (2.9) 0 (0.0) 6 1 (1.0) 1 (100.0) Total Values are percents of total number of patients. b Values are percents of patients with specified number of ovulations. 298 Silverberg et al. Follicular size predicts ovulation

4 Table 3 Ovulation and Cycle Outcome Conception (n = 19) Nonconception (n = 84) Total (n = 103) Probability No. of follicles No. of ovulations No. of follicles > 15 mm No. of ovulations > 15 mm 11.7 ± ± ± ± ± ± ± ± ± 0.2 b Values are means ± SEM. b, not significant. multiple ovulation did not ensure multiple gestation. All six multiple gestations resulted from cycles in which multiple ovulations occurred; however, four of the six singleton gestations also resulted from multiple ovulation cycles. Patients presented for consecutive cycles of controlled ovarian hyperstimulation on 54 occasions. All were evaluated for the presence of residual follicular cysts before reinitiating controlled ovarian hyperstimulation. Cysts were found in 28 patients (51.9%). Serum E 2 levels on the day ofhcg administration correlated with the presence of residual cysts (r = 0.795, P < 10-8 ). However, there was no correlation between the presence of residual ovarian cysts and either the number of ovulatory follicles or the total number of follicles present on the day of hcg administration. In addition, there was no relationship between the presence of residual cysts and either the total hmg dose administered or patient diagnosis. DISCUSSION Numerous hmg regimens and recommendations for hcg timing have been proposed for controlled ovarian hyperstimulation. 6 8 Several reasons account for the lack of a consensus on these issues. First, oocyte maturity has never been correlated with follicular size in hmg-stimulated controlled ovarian hyperstimulation cycles. This has led to an empirical approach to hcg timing, with some advocating administration at smaller follicular sizes to prevent oocyte atresia and others recommending later administration to ensure oocyte maturity. 9 Second, follicular size has never been correlated with the incidence of ovulation in these cycles. When a follicle attains a specific size, there is still no guarantee that it will ovulate. In fact, the incidence of ovulation at any given follicular size has heretofore never been determined in hmg cycles. Third, no specific number of ovulated oocytes leading to an increased likelihood of conception has been identified, and fourth, no optimal serum E 2 range correlating with increased fecundity has been determined. Because our lack of knowledge in these areas hampers the development of protocols leading to improved cycle outcome, some recommendations have been made in an attempt to reduce the likelihood of adverse outcomes, specifically, multiple gestations and ovarian hyperstimulation syndrome With the advent of improved cycle-monitoring techniques, the incidence of these adverse outcomes has declined. 11 By systematically addressing the already enumerated points, it should also be possible to increase the number of favorable outcomes. This study was designed to address the second point, i.e., the correlation between follicular size and ovulation. In natural cycles, a follicle attains a mean diameter of 20 to 25 mm before ovulation. 2 Because the majority of women receiving hmg stimulation will not ovulate spontaneously, it becomes more difficult to ascertain periovulatory follicular dimensions in these cycles. 12 If we assume that only mature oocytes are capable of ovulation, it appears, based on ovulation rates, that oocyte maturity may be achieved when a follicle reaches a mean diameter of 15 to 16 mm. This observation is consistent with studies demonstrating IVF pregnancies achieved with oocytes obtained from follicles as small as 16 mm. 7 The critical size breakpoint for ovulation, however, appears to be 17 mm, as follicles 17 mm or larger Table 4 Comparison of Ovulatory Follicle Size in Conception and Nonconception Cycles No. of cycles with ovulation of follicles measuring :S14 mm 15 to 16 mm 17 to 18 mm 19 to 20 mm >20mm Conception (n = 19) 1 (5.3)b 8 (42.1) 10 (52.6) 10 (52.6) 4 (21.1) Nonconception (n = 84) 3 (3.6) 37 (44.0) 54 (64.3) 32 (38.1) 13 (15.5) None of the conception cycle values differs significantly from the corresponding nonconception cycle value. b Values in parentheses are percents. Vol. 56, No.2, August 1991 Silverberg et al. Follicular size predicts ovulation 299

5 have a 13-fold higher incidence of ovulation than do those 16 mm or smaller. Although the vast majority of follicles that develop during controlled ovarian hyperstimulation are <15 mm on the day of hcg administration, these follicles rarely ovulate. They do, however, contribute to total serum E 2 levels and may predispose to the development of ovarian hyperstimulation syndrome. 13 These data demonstrate that follicular size on the day of hcg administration does correlate with the incidence of ovulation, and this enabled us to derive the ovulation prediction equation. Using multiple linear regression analysis, the data collected in this study permitted derivation of an equation to determine the relative contribution of any given follicle to total serum E 2 level. All anovulatory patients were World Health Organization group II normoestrogenic women. 14 These women, like normal ovulatory patients, tend to develop multiple dominant follicles in response to hmg stimulation. Several studies have demonstrated a poor correlation between serum E 2 level and the dim ensions of the dominant follicle(s) in these patients However, no studies have determined the contribution of individual follicles to total E 2 levels. Multiple linear regression analysis results suggest that follicles produce low levels of E 2 until they reach 14 mm. Then, there is a marked, progressive increase in E 2 production. When follicles reach 17 mm, continued growth results in little incremental E 2 production. An optimal serum E 2 range or an optimal number of ovulatory follicles that correlates with conception was not present in this study. The latter observation suggests that a larger number of ovulatory follicles does not necessarily increase the chance for conception. Our statistical power for these observations is low, however, and it remains possible that a larger number of pregnancies would enable determination of optimal ranges for these parameters. In summary, ovulation rates are dependent on follicle size at the time of hcg administration. In addition, serum E 2 levels can be predicted based on follicle size and number. The derived ovulation prediction equation provides the first objective criterion correlating ovulation with follicle size on the day of hcg administration in hmg-stimulated cycles. The regression equation to predict E 2 values may have clinical utility, by obviating the need for costly daily serum E 2 determinations. Should future studies define an optimal number of ovulations or range of serum E 2 values that correlate with conception, these equations will assume an even larger role in patient management. Acknowledgment. We gratefully acknowledge Mrs. Sylvia Crisantes for the diligent preparation of this manuscript. REFERENCES 1. Blankstein J, Shalev J, Saadon T, Kukia EE, Rabinovici J, Pariente C, Lunenfeld B, Serr DM, Mashiach S: Ovarian hyperstimulation syndrome: prediction by number and size of preovulatory ovarian follicles. Fertil Steril47:597, Queenan JT, O'Brien GD, Bains LM, Simpson J, Collins WP, Campbell S: Ultrasound scanning of ovaries to detect ovulation in women. Fertil Steril 34:99, Seibel MM, McArdle C, Smith D, Taymor ML: Ovulation induction in polycystic ovarian syndrome with urinary folliclestimulating hormone or human menopausal gonadotropin. Fertil Steril 43:703, Haning RV, Jr, Austin CW, Carlson IH, Kuzma DL, Shapiro SS, Zweibel WJ: Plasma estradiol is superior to ultrasound and urinary estriol glucuronide as a predictor of ovarian hyperstimulation during induction of ovulation with menotropins. Fertil Steril 40:31, Ritchie WGM: Ultrasound in the evaluation of normal and induced ovulation. Fertil Steril43:167, March CM: Human menopausal gonadotropins. In Infertility and Reproductive Medicine Clinics of North America, Edited by B Yee. Philadelphia, W.B. Saunders, 1990, p Trounson AO: In vitro fertilization. In Current Topics in Experimental Endocrinology: Pregnancy and Parturition, Vol. 5, Edited by L Martini, V James. New York, Academic Press, 1983, p Collins JA: Superovulation in the treatment of unexplained infertility. Semin Reprod Endocrinol 8:165, Williams RF, Hodgen GD: Disparate effects of human chorionic gonadotropin during the late follicular phase in monkeys: normal ovulation, follicular atresia, ovarian acyclicity, and hypersecretion of follicle-stimulating hormone. Fertil Steril 33:64, March CM: Complications of gonadotropin therapy. J Reprod Med 21:208, March CM: Improved pregnancy rate with monitoring of gonadotropin therapy by three modalities. Am J Obstet Gynecol 156:1473, Talbert LM: Endogenous luteinizing hormone surge and superovulation. Fertil Steril 49:24, Blankstein J, Shalev J, Saadon T, Kukia EE, Rabinovici J, Pariente C, Lunenfeld B, Serr DM, Mashiach S: Ovarian hyperstimulation syndrome: prediction by number and size of preovulatory ovarian follicles. Fertil Steril 47:597, Ho Yuen B, PrideS: Induction of ovulation with exogenous gonadotropins in anovulatory infertile women. Semin Reprod Endocrinol 8:186, Haning RV, Jr, Austin CW, Kuzma DL, Shapiro SS, Zweibel WJ: Ultrasound evaluation of estrogen monitoring for induction of ovulation with menotropins. Fertil Steril 37:627, Sallam HN, Marinho AO, Collins WP, Rodeck CH, Campbell S: Monitoring gonadotropin therapy by real-time ultrasonic scanning of ovarian follicles. Br J Obstet Gynaecol 89:155, Silverberg et al. Follicular size predicts ovulation