The periovulatory and luteal phase of conception cycles following in vitro fertilization and embryo transfer

Transcription

1 FERTILITY AND STERILITY Copyright 1984 The American Fertility Society Prinl d in U.8A. The periovulatory and luteal phase of conception cycles following in vitro fertilization and embryo transfer Alexander M. Dlugi, M.D. * Neri Laufer, M.D.t Alan H. DeCherney, M.D. Neil J. MacLusky, Ph.D. Florence P. Haseltine, Ph.D., M.D. Mary Lake Polan, M.D., Ph.D. Howard C. Mezer, M.D. Basil Tarlatzis, M.D. Frederick Naftolin, M.D., D.Phil. Department of Obstetrics and Gynecology, Yale University School of Medicine, New Haven, Connecticut The pattern of periovulatory and luteal phase levels of serum estradiol (E,J and progesterone (P) were compared between 8 conception and 28 nonconception cycles of patients undergoing in vitro fertilization (lvf). Ten additional women served as control subjects and did not undergo follicular aspiration. Follicle growth was induced with an individualized Pergonal (human menopausal gonadotropin) regimen, and laparoscopy was performed 36 hours after human chorionic gonadotropin administration. The length of the luteal phase did not differ significantly among the three groups and was between 14 and 15 days in duration. When IVF conception cycles were compared with nonconception cycles, although no difference in the number of large follicles was observed (4.25 ± 0.45 versus 3.6 ± 0.25), the patterns of E2 and P differed significantly. Daily serum E2 levels tended to be higher in the periovulatory phase in conception cycles when compared with nonconception cycles, and were significantly (P < 0.05) higher in the early, mid, and late luteal phases. Serum P levels were significantly higher (P < 0.05) in conception cycles from the midluteal phase onward. A decline in both serum E2 and P in the midluteal phase in conception cycles suggested some degree of corpus luteum deficiency. It is suggested that high E2 levels in the periovulatory phase may be an indicator of better follicular development under human menopausal gonadotropin stimulation and that the deficiency observed in the late luteal phase is overcome with the establishment of pregnancy. Fertil Steril 41 :530, 1984 The initial pregnancies obtained through in vitro fertilization (IVF) in humans were achieved with oocytes from unstimulated ovaries. 1 Because Received September 16, 1983; revised and accepted December 28, *Reprint requests: Alexander M. Dlugi, M.D., Department of Obstetrics and Gynecology, Yale University School ofmedicine, 333 Cedar Street, P.O. Box 3333, New Haven, Connecticut t Andrew W. Mellon Foundation Fellow in Reproductive Sciences on leave from the Department of Obstetrics and Gynecology, Hadassah University Hospitlll, Jerusalem, Israel. 530 Dlugi et a1. Luteal phase of conception cycles in NF the yield in the natural cycle is low, and because pregnancy rates have improved with multiple embryo transfer (ET), most groups involved with IVF now use cycles stimulated with clomiphene or a combination of clomiphene and human menopausal gonadotropin (hmg).2 Two main factors play a major role in the success of IVF: (1) the quality of the egg and the ensuing embryo after fertilization; and (2) the adequacy of the luteal phase and its effect in establishing a uterine environment suitable for nidation. In the normal ovulatory cycle, the effect of Fertility and Sterility

2 follicular aspiration on the luteal phase has been controversial. Kerin et al.,3 Kemeter et al.,4 and Frydman et al. 5 demonstrated no effect of follicle aspiration on either the length of the luteal phase or the levels of measured steroid hormones. On the other hand, Jones et al. 6 noted a small but significant decrease in both progesterone (P) and estradiol (E2) levels in aspirated cycles where an oocyte was obtained but ET was not accomplished. In clomiphene-induced cycles where oocytes were aspirated for IVF, both Frydman et al. 5 and Kemeter et al.4 found no effect of aspiration on the luteal phase. hmg as a sole drug for ovarian stimulation for IVF was initially pioneered by' Edwards et al. 7 but later abandoned. These authors noted shortened luteal phase in all patients treated with hmg. Recently, only Jones et al. 8 and Laufer et al. 9 reported the use of moderate and high doses of hmg, respectively, for stimulating follicular growth for IVF. Jones et al. 6 reported that the luteal phase in their first two successful pregnancies was characterized by a high serum P value but with a fall that indicated that the luteal phase was destined to be short except for the intervention of the pregnancy. Because pregnancy is the ultimate proof of successful IVF and ET in terms of corpus luteum function, it was the purpose ofthis study to retrospectively compare the patterns of periovulatory and luteal phase levels of serum E2 and P in failed and successful cycles of IVF patients as well as in patients who underwent similar induction of ovulation without follicle aspiration. PATIENT SELECTION MATERIALS AND METHODS Three groups of patients were retrospectively compared. The first group (group I) consisted of ten normally ovulating women with at least one patent fallopian tube who underwent induction of ovulation with hmg after failing to conceive at least 18 months following microsurgical tubal reconstruction; this group did not undergo IVF and follicle aspiration and thus served as a control population. Thirty-six women were chosen from a larger group of 50 women participating in IVF between September 25, 1982, and January 5, 1983, on the basis of the completeness of the laboratory data at the time of this writing. Group II consisted of 28 women who failed to conceive following ET. Eight women who conceived following the transfer of two to eight fertilized oocytes composed the third group (group III) of this study. Eight women from group II were selected on the basis of having a similar number of embryos transferred as the eight women conceiving and constitute the subgroup lia. The patients' ages in all three groups ranged from 25 to 37 years, with a mean of 32.5 years. All women participating suffered from primary or secondary infertility of 2 to 10 years' duration (mean, 5.5 years) as a result of tubal pathology. All had normal ovulatory function and were euprolactinemic. The minimal criteria for inclusion in the IVF program were previously reported. 10 OVULATION INDUCTION AND MONITORING SYSTEM The patients received three ampules of hmg (Pergonal, Serono Laboratories, Randolph, MA) (225 IU/day of follicle-stimulating hormone and luteinizing hormone) from day 3 of the cycle for 5 days. Monitoring consisted of daily serum E2 and ovarian ultrasonography (PHO-Sonic B-scanner, Siemens Medical Systems, Iselin, NJ) starting on day 8 of the cycle. No measurements of serum luteinizing hormone were performed. The criteria for adequate follicular growth and human chorionic gonadotropin (he G) administration (10,000 IU administered intramuscularly) were an E2 level> 400 pg/ml and at least two follicles with a diameter> 1.6 cm. If these criteria were not met by day 8, the hmg dose was increased by an additional one to two ampules/day for 1 to 4 days. A range of 15 to 33 ampules/cycle ofhmg (mean, 19 ± 4) was administered to these women. The day of heg administration is designated day o. LAPAROSCOPY AND ASPIRATION OF FOLLICLES Laparoscopy was performed in the morning 36 to 38 hours after heg administration. Ovarian ultrasonography was performed 30 minutes to 1 hour prior to the procedure; and if fewer than two large follicles were visualized, the procedure was abandoned. The oocytes were aspirated with the use of a three-puncture technique.l1 The needle used for the aspiration was 13-gauge with an introducer modified from the TRU-cut biopsy needle (2N-27-04, Travenol Laboratories, Deerfield, IL). The needle was washed prior to aspiration with heparinized (2 ng/md medium devoid of serum. A Dlugi et al. Luteal phase of conception cycles in NF 531

3 20-ml DeLee suction trap (Argyle Company, St. Louis, MO) was connected to the aspirating needle by an extension tube connected to a continuous wall suction at 100 mm Hg. The ovaries were grasped by the ovarian ligament for fixation, and the follicles entered at an avascular area. An average of 3 ml of follicular fluid was aspirated. When blood-tinged follicular fluid was obtained, an equal volume of warm heparinized medium was added to the trap. No attempt was made to flush the aspirated follicles. The collection traps were rushed to the laboratory in a Styrofoam container cushioned with prewarmed saline infusion bags. The laboratory was located 5 to 7 minutes' walking distance from the operating theater. CULTURE MEDIA AND FERTILIZATION The details have been previously reported. 9 Briefly, modified Ham's F-10 medium was prepared weekly from a stock solution as described by Lopata et al. ll The insemination medium (1M) was supplemented with 10% of the individual woman's heat-inactivated serum. Growth medium (GM) was similarly prepared to contain 20% serum. Ova and embryos were cultured in 1 ml medium in organ culture dishes (Falcon #3037, Falcon Plastics, Oxnard, CA). The medium was incubated overnight at 37.5 C in sealed jars after gassing with 5% CO 2, 5% O 2, and 90% N 2 The length of preincubation prior to insemination was determined individually for each oocyte. This decision depended on the degree of cumulus modification and corona cell dispersion present Immature oocyte/corona/cumulus complexes were preincubated for 24 hours prior to insemination, and more mature complexes were preincubated for 6 to 8 hours before insemination. Eggs were inseminated with 1 x 10 6 motile sperm in 25 to 50 f.li 1M. The ova were moved from insemination medium to growth medium 16 to 18 hours after insemination and denuded mechanically of excess corona cells by repeated pipetting. These oocytes were examined for the formation of pronuclei. At 38 to 40 hours after insemination the 2- to 4-cell embryos were implanted in a volume not exceeding 80 f.li GM with the use of an 18- to 20-gauge Teflon-tipped catheter with a side opening. The women were implanted while they were in the knee-chest position if the uterus was anteverted and in the dorsal lithotomy position of the uterus was retroverted. 532 Dlugi et al. Luteal phase of conception cycles in NF TREATMENT AND MONITORING AFTER IMPLANTATION Patients undergoing IVF were hospitalized for 24 hours and kept at bed rest. Intramuscular P, 25 mg/day, was begun on the day of implantation and continued throughout the luteal phase. Control patients received no hormonal supplementation. Blood was drawn every 3 days after the hcg injection for ~-hcg titer, P, and E 2 Conception was detected by a rising l3-hcg titer and followed up to 10 to 12 weeks of gestation. Ultrasonography was done weekly until 10 to 12 weeks' gestation and biweekly thereafter and confirmed the presence of an intrauterine gestation. STATISTICS The results were analyzed using analysis of variance (ANOV A), Student's t-test, and Duncan's Multiple Range test. The ANOV A and Student's t-test calculations were performed on an Apple II microcomputer (Apple Computer Inc., Cupertino, CA), using programs obtained from Human Systems Dynamics, Northridge, CA. The Duncan tests were performed with a Texas Instruments Ti 59 programmable calculator, with a program obtained through ppx-59 (Texas Instruments, Lubbock, TX). Data are presented throughout this paper as the mean ± standard' error of the mean (SEM). RESULTS LENGTH OF THE LUTEAL PHASE The length of the luteal phase (15.8 ± 0.54 days) in those patients who underwent follicle aspiration and who received P supplementation but who did not conceive following IVF and ET (group II) did not differ significantly from that of the control population (14.4 ± 0.5 days) which did not receive hormonal supplementation (group I). NUMBER OF FOLLICLES The mean number of large (> 1.5 cm in diameter) follicles visualized on ultrasonography in each of the three groups is shown in Table 1. The group that conceived (group III) had a significantly (P < 0.05) greater number offollicles (4.25 ± 0.45) as compared with the control group (2.75 ± 0.25) (group I) did not differ from group II (3.6 ± 0.25). At times an individual aspirated follicle yielded more than one oocyte and accounts for the Fertility and Sterility

4 Table 1. Clinical Data on the Study Poputuliurt No. of em- Group No. of bryos trans- Le:f!h of follicles ferred lute phase days I 2.75 ± ± 0.5 II 3.6 ± ± ± lI 4.25 ± 0.45 a 5.13 ± 0.66 b ap < 0.05, group III versus group I (Student's t-test). bp < 0.001, group III versus group II (Student's t-test). 8 7 E2 6 In Pg/ml 5 ~ Non- Pregnant o Pregnant discrepancy in group III between the number of aspirated follicles and the number of transferred embryos. NUMBER OF EMBRYOS TRANSFERRED The number of embryos transferred to those patients in the pregnant group (5.13 ± 0.66) differed significantly (P < 0.001) from the numbertransferred to those patients in the nonpregnant group (2.78 ± 0.29) E2 800 Pg/ml Day of Cycle Figure 1 Pattern of serum E2 levels in control (~), nonconception (0), and conception (e) cycles. Day 0 represents the day of hcg administration. Not all patients had blood drawn on every given day. The number of serum samples measured per day of the cycle in group I ranged from 5 to 10 but on days - 2 through 0 ranged from 8 to 10. In group II, the sample size ranged from 9 to 28 but on days -1 through + 14 ranged from 25 to 28. In group III, the sample size ranged from 5 to 8 but on days -1 through + 14 ranged from 7 to 8. Serum E2 levels are significantly higher (*P < 0.05; Duncan's Multiple Range test) in conception cycles on days + 1, + 8, + 14, and + 17, as compared with nonconception cycles I o Days Before HCG Administration Figure 2 A comparison of periovulatory serum E2 levels between conception and nonconception cycles undergoing follicle aspiration and ET. Day 0 represents the day ofhcg administration. Although conception cycles tend to demonstrate higher E2 levels, the difference is not significant. ESTRADIOL The pattern of serum E2 levels in the periovulatory and luteal phases in all three groups is shown in Figure 1. The pattern is similar in all three groups, with two peaks, the first occurring at day 0 to + 1, followed by a gradual decline over days + 2 to + 5; the second peak occurred on day + 8, followed by a second decline toward day + 12 (group II nadir, ± 111.9, range, 39 to 312 pglml; group III nadir, 336 ± 84.6, range, 83 to 763 pg/ml). In group III, however, declining E2 levels began to increase with the establishment of pregnancy. In the periovulatory phase, control (group I) cycles showed a mean peak E2 level of 656 ± 91 pg/ml (range, 339 to 1559 pg/md on the day of hcg administration, followed by a decrease; whereas stimulated cycles undergoing NF showed a continued rise to day + 1 (group II mean,861.9 ± 107.2, range, 236 to 2194 pg/ml; group III mean, ± 283.2, range, 406 to 2669 pg/ml), which then declined from day + 2 onward. Daily comparisons among the three groups showed no difference between group I and group II, whereas group III periovulatory E2 levels tended to be higher from day - 2 up to the day ofhcg administration, although not significantly (Fig. 2). Of the women who subsequently conceived, 62.5% achieved an E 2 level > 900 pg/ml on day 0, whereas only 35.7% of women not conceiving attained that level. Group III patients, those conceiving following IVF, showed significantly (P Dlugi et ai. Luteal phase of conceptwn cycles in IVF 533

5 Table 2. Comparison of Serum E2 Levels in the Periovuiatory and Luteal Phases of Conception and Nonconception Cycles Matched for the Number of Embryos Transferred" Group Day of cycle " +17 pglml ITa ± 76 ± 79 ± 61 ± 202 ± 102 ±101 ± 115 ± 30 ±6 ±2 ill b b 1446 b c 759 c ±109 ± 174 ±143 ± 283 ± 305 ± 179 ± 307 ± 79 ± 213 ± 182 "The number of serum samples measured per group per day of the cycle ranged from 2 to 8; however, on days -1 through + 14, the number of samples measured ranged from 7 to 8. Statistical analysis, two-way ANOV A: P < 0.001, day of cycle, F = 5.5, df = 9.119; P < 0.001, patient group, F = 30.7, df = bp < 0.05, Student's t-test. cp < 0.01, Student's t-test. < 0.05) higher E2 levels on the day preceding laparoscopy (day + 1), in the midluteal phase (day + 8), and following the establishment of pregnancy (day + 14 and onward). In order to correct for the possible effect of increased numbers of embryos transferred on E2 levels in conception cycles, and in order to gain better insight into follicular determinants of success in IVF, a further comparison was made between pregnant cycles (group III) and eight women who were implanted with the same number of embryos (group IIa). For each patient conceiving, one patient in group IIa was compared. The mean number of embryos transferred to the patients in group IIa was 4.5 ± Although the basic pattern remains unchanged, conceptual cycles now demonstrate additional significantly (P < 0.05) higher E 2 levels on days 0 and + 5 (Table 2). PROGESTERONE The pattern of serum P levels in the periovulatory and luteal phases in all three groups is shown in Figure 3. Small concentrations of P were detected in the periovulatory phase from day - 2 up to the day ofhcg administration in all groups, with levels ranging from 0.4 to 2.4 ng/ml. From day 0 onward, a progressive increase in P levels was observed, peaking on day + 8, and then gradually decreasing in all three groups toward day + 12 (group II nadir, 17.4 ± 2.0, range, 4.3 to 51.2 ng/ml; group III nadir, 36.0 ± 9.1, range, 8.5 to 83 ng/ml). Similar to E2 levels, declining P levels in group III began to increase with the establishment of pregnancy. P levels in group I did not differ from those of group II throughout the luteal phase. However, serum P levels in conceptual cycles (group III) differed markedly from 534 D1ugi et ai. Luteal phase of conception cycles in IVF the midluteal phase onward. When group IIa (matched for number of embryos transferred) was compared with group ill, there was no difference in midluteal P; but with the establishment of pregnancy, late luteal P levels differed significantly (Table 3). PROGESTERONEIESTRADIOL RATIO PlE 2 ratios tended to be low for all three groups up to day 0 but then gradually increased. In group I, the values from the early (day + 3) luteal phase onward throughout the mid and late luteal phasi. es remained level at In both groups (groups so 70 j' * *,,, ",* 60 I,, Progesterone 50 t/ \ I ng/ml IJ. \*: I /1 \ :! Group m 0~~~L-L-L-~L-L-L-~ Group IT Figure 3 Pattern of serum P levels in control (~), nonconception (0), and conception (e) cycles. Day 0 represents the day!>fhcg administration. Not all patients had blood drawn on every given day. The number of serum samples measured per day of the cycle in group I ranged from 3 to 7 but on days + 5 through + 8, 7 were measured. In group IT, the sample size ranged from 4 to 28 but on days + 5 through + 14 ranged from 23 to 28. In group ITI, the sample size ranged from 2 to 8 but on days + 2 through + 14 ranged from 7 to 8. Serum P levels are significantly higher (*P < 0.05; Duncan's Multiple Range test) in conception cycles on days +8, +12, +14, and +17, as compared with both group I and group II. Fertility and Sterility

6 Table 3. Comparison of Serum P Levels in the Periovulatory and Luteal Phases of Conception and Nonconception Cycles Matched for the Number of Embryos Transferred" Group Day of cycle IIa 0.98 ± ± ± ± ± ± ± 1.6 m 1.67 ± ± ± ± ± ± b ± c ± ± 10.5 "The number of serum samples measured per group per day of the cycle ranged from 1 to 8; however, on days +2 through + 14, the number of samples measured ranged from 7 to 8. Statistical analysis (days -1 to + 14), two-way ANOV A: P < 0.001, day of cycle, F = 9.2, df = 7.77; P < 0.02, patient group, F = 6.5, df = bp < 0.05, Student's t-test. cp < 0.01, Student's t-test. nglml II and III) treated with P on the day of ET (day + 4), the P/E2 level was at least twice that of control subjects throughout the remaining luteal phase. However, because of the large standard deviations, no differences between these three populations were found. DISCUSSION Specific single factors responsible for the success of IVF have yet to be determined. It would appear that follicular recruitment and oocyte maturation playa significant role. In the periovulatory phase, the data reported here demonstrate higher levels of E2 in conception cycles. Indeed, 62.5% of women in this study who subsequently conceived achieved an E 2 level > 900 pg/ml on the day of hcg administration (day 0), whereas only 35.7% of women not conceiving attained that level. This difference cannot be attributed to increased follicular recruitment, because the number of large follicles in conception cycles (group III) was no different from that in nonconception cycles (group 11). Furthermore, when specifically controlled for the number of embryos transferred, conception cycles continued to demonstrate this trend of higher E2 levels, reaching significance prior to the time ofhcg administration. This may reflect a more mature cohort of follicles, irrespective of the number, that is capable of producing higher E2 levels. This group may be analogous to the high E2 responders reported by Garcia et al.13 These authors suggested that a higher than normal E2 response to hmg-stimulated cycles might be required for full follicular development and oocyte maturation to occur. However, it is not clear that conception rates are improved. Nonetheless, the elevated serum E2 levels may reflect a better steroidogenic capacity of the more mature follicles in conception cycles. In this study the finding that high periovulatory E2 levels correlated well with subsequent conception may also be a reflection of the importance of a high estrogenic milieu for oocyte maturation. The importance of high intrafollicular estrogen levels has been documented in animal models. In the sheep, rabbit, and human, complete oocyte maturation, including both cytoplasmic and nuclear maturation, is probably regulated by both E2 and gonadotropins.14 This may be the basis for the finding, previously reported,15 that fertilized oocytes showed increased follicular levels of E 2 A high intrafollicular estrogenic milieu is important for oocyte maturation in the human, because it was shown that these follicles are the sole source of oocytes capable of fertilization and continued development into a successful pregnancy!s,17 Although oocyte maturation appears to be a crucial factor, the in vitro process itself may compensate for incomplete in vivo oocyte development. Recently, Veeck et al.10 reported the establishment of two pregnancies resulting from the transfer of oocytes matured in vitro. This would suggest that oocyte maturation in vivo may not be an absolute necessity in terms of IVF success. In contrast to other reports,s, 7 the length of the luteal phase, its pattern, and the serum levels of both E2 and P appeared to be unaffected by follicle aspiration (group IT or group I). Both control and nonconception cycles demonstrate a luteal phase pattern similar to that described by Laufer. et al. 18 for nonfertile, nonaspirated cycles. One might be tempted to conclude, therefore, that aspiration does not affect the luteal phase. Group IT patients, however, were supplemented with parenteral P, and although serum P levels and luteal phase length were not different as compared with control subjects, a biologic effect on cervical mucus was noted, and the PlE 2 ratio was at least twice that of control subjects. One may speculate Dlugi et al. Luteal phase of conception cycles in NF 535

7 that P supplementation may have masked or corrected a potential corpus luteum deficiency. Pregnant cycles (group III) demonstrated a sharp decline of both E2 and P from day + 8 to day + 12, with a subsequent rescue with the establishment of pregnancy. This confirms the data of Jones et ai.,8 who reported two conception cycles with a luteal phase that was characterized by a high P value but with a fall that indicated that the luteal phase was destined to be short except for the intervention of a pregnancy. This is in sharp contrast, however, to the work of Laufer et ai.,18 who reported a continuous rise of both E2 and P in the late luteal phase in both spontaneous and induced fertile, nonaspirated cycles. One cannot ascribe this difference to the removal of large numbers of granulosa cells as a result of the aspiration process in in vitro cycles. The number of granulosa cells aspirated per follicle ranges from 1 x 10 5 to 2 X 10 6,19 constituting < 10% of the total granulosa cell cohort, as calculated by McNatty and co-workers20 for mature preovulatory follicles. Other factors, then, must be sought. to explain this relative corpus luteum deficiency. Edwards et ai. 7 noted an inverse relationship between estrogens excreted during the follicular phase and the length of the luteal phase in induced cycles and suggested a possible luteolytic effect of high estrogen levels. On the other hand, E2 levels in the induced, fertile, and nonfertile nonaspirated cycles reported by Laufer et ai. 18 were significantly higher than those in cycles with spontaneous ovulation; yet no corpus luteum deficiency was detected. Hyperprolactinemia has been associated with luteal phase defects.21 High prolactin (PRL) levels have been found in women undergoing ovarian stimulation with either hmg12 or human pituitary gonadotropin22; the increase in PRL occurred in parallel with an increase in E2 in these studies. It was suggested that because estrogens are potent stimulators of PRL secretion, the high serum E2 levels found during hmg treatment may be the cause of the corresptmdingly high serum PRL levels. Alternatively, the effect of anesthesia prior to ova aspiration must be considered: under all types of anesthesia, a highly significant rise in serum PRL occurs.23 Although serum PRL levels were not measured in this study, it is possible that a relative hyperprolactinemic state may have contributed to a relative corpus luteum deficiency. Finally, Heap et ai.24 have reviewed some 536 D1ugi et al. Luteal phase of conception cycles in IVF of the data suggesting a role for early embryoni signals in the initiation of maternal recognition of pregnancy prior to implantation. Thus, chori onic gonadotropin in rhesus monkeys, trophoblas tin in sheep, and estrogen in pigs have been pos tulated to exert either antiluteolytic or luteotro phic effects. No specific early embryonic signa has been identified in humans, although early pregnancy factor may have a potential role. Jone et ai.8 detected hcg some 13 days after follicle aspiration. This rise in hcg appears to rescue the failing corpus luteum, serum P and E2 having declined earlier. If other early embryonic factor do indeed exist in the human, one may speculat that the IVF process may in some way interfer with their signal, perhaps through delayed pro duction, and thus result in a failure of materna recognition of pregnancy with a subsequent de cline in corpus luteum function. It is concluded that serum E21evels in the peri ovulatory phase tend to be higher in conception cycles when compared with nonconceptual cycle in IVF and seem to be an indicator of better fol licular development under hmg stimulation. Lu teal phase serum E2 and P in conception cycle seem to show a decline in the latter part of thi phase, probably due to some degree of corpus lu teum insufficiency. This deficiency may be over come by parenteral administration of P and i rescued by the establishment of pregnancy. Acknowledgment. We are indebted to Dr. Bennett Shaywit of the Department of Pediatrics, Yale University School o Medicine, for use of the Apple II statistical programs. REFERENCES 1. Steptoe P, Edwards R: Birth after the implantation of human embryo. Lancet 2:366, Edwards RG, Purdy JM (Eds): Patient selection and mon itoring. In Human Conception in Vitro. London, Academ ic Press, 1982, p Kerin JF, Broom TJ, Ralph MM, Edmonds DK, Warne GM, Jeffrey R: Human luteal phase function followin oocyte aspiration from the immediately preovular graaf ian follicle of spontaneous ovular cycles. Br J Obstet Gy naecol 88:1021, Kemeter P, Feichtinger W, Neurriark J, Szalay S, Biegl mayer Ch, Janisch H: Influence of laparoscopic follicula aspiration under general anaesthesia on corpus luteum progesterone secretion in normal and clomiphene-stimu lated cycles. Br J Obstet Gynaecol 89:948, Frydman R, Testart J, Giacomini P, Imbert MC, Marti E, Nahoul K: Hormonal and histological study of the lu teal phase in women following aspiration of the preovula tory follicle. Fertil Steril 38:312, 1982 Fertility and Sterili

8 6. Jones GS, Garcia J, Acosta A: Luteal phase evaluation in in vitro fertilization. In Human Conception in Vitro, Edited by RG Edwards, JM Purdy. London, Academic Press, 1982, p Edwards RG, Steptoe PC, Purdy JM: Establishing fullterm human pregnancies using cleaving embryos grown in vitro. Br J Obstet Gynaecol 87:737, Jones HW Jr, Jones GS, Andrews MC, Acosta A, Bundren C, Garcia J, Sandow B, Veeck L, Wilkes C, Witmyer J, Wortham JE, Wright G: The program for in vitro fertilization at Norfolk. Fertil Steril 38:14, Laufer N, DeCherney AH, Haseltine FP, Polan ML, Mezer HC, Dlugi AM, Sweeney D, Nero F, Naftolin F: The use of high-dose human menopausal gonadotropin in an in vitro fertilization program. Fertil Steril 40:734, Veeck LL, Wortham JWE Jr, Witmyer J, Sandow BA, Acosta AA, Garcia JE, Jones GS, Jones HW Jr: Maturation and fertilization of morphologically immature human oocytes in a program of in vitro fertilization. Fertil Steril 39:594, Lopata A, Johnston IWH, Hoult IJ, Speirs AL: Pregnancy following intrauterine implantation of an embryo obtained by in vitro fertilization of a preovulatory egg. Fertil Steril33:117, Yuen BH, McComb P, Sy L, Lewis J, Cannon W: Plasma, prolactin, human chorionic gonadotropin, estradiol, testosterone and progesterone in the ovarian hyperstimulation syndrome. Am J Obstet Gynecol 133:316, Garcia J, Jones GS, Acosta AA, Wright GL Jr: Corpus luteum function after follicle aspiration for oocyte retrieval. Fertil Steril 36:565, Thibault C: Are follicular maturation and oocyte maturation independent processes? J Reprod Fertil 51:1, Carson R, Trounson A, Findlay J: Successful fertilization of human oocytes in vitro: concentration of estradiol-17j3, progesterone and androstenedione in the antral fluid of donor follicles. J Clin Endocrinol Metab 55:798, Laufer N, Botero-Ruiz W, DeCherney AH, Haseltine FP, Polan ML, Behrman HR: Gonadotropin and prolactin levels in follicular fluid of successfully fertilized human ova in vitro. J Clin Endocrinol Metab. In press 17. Botero-Ruiz W, Laufer N, DeCherney AH, Polan ML, Haseltine FP, Behrman HR: The relationship between follicular fluid steroid concentration and successful fertilization of human oocytes in vitro. Am J Obstet Gynecol. In press 18. Laufer N, Navot D, Schenker JG: The pattern of luteal phase plasma progesterone and estradiol in fertile cycles. Am J Obstet Gynecol 143:808, Polan ML, Ohkawa R, Botero-Ruiz W, Laufer N: Direct effect of prolactin on progestin production by cultured granulosa cells from preovulatory human follicle. Presented at the Sixty-Fifth Annual Meeting of the Endocrine Society, June 8 to 10, 1983, San Antonio, Texas, p McNatty K, Sawers R, McNeilly A: A possible role for prolactin in control of steroid secretion of the human graafian follicle. Nature 250:653, Andrews WC: Luteal phase defects. Fertil Steril 32:501, Healy D, Burger H: Serum follicle-stimulating hormone, luteinizing hormone and prolactin during induction of ovulation with exogenous gonadotropins. J Clin Endocrinol Metab 56:474, Szalay S, Feichtinger W, Kemeter P, Beck H, Janisch H, Neumark J: Changes in hormonal parameters under different kinds of general anesthetics during laparoscopy oocyte recovery. In Human Conception in Vitro, Edited by RG Edwards, JM Purdy. London, Academic Press, 1982, p Heap R, Flint A, Gadsby J: Embryonic signals and maternal recognition. In Cellular and Molecular Aspects of Implantation, Edited by SR Glasser, DW Bullock. New York, Plenum Publishers, 1981, p 311 Dlugi et ai. Luteal phase of conception cycles in IVF 537