K.W.Fuh, X.Wang, A.Tai, I.Wong and R.J.Norman 1

Human Reproduction vol.12 no.10 pp.2162 2166, 1997 Intrauterine insemination: effect of the temporal relationship between the luteinizing hormone surge, human chorionic gonadotrophin administration and insemination on pregnancy rates K.W.Fuh, X.Wang, A.Tai, I.Wong and R.J.Norman 1 Reproductive Medicine Unit, Dept of Obstetrics and Gynaecology, University of Adelaide, The Queen Elizabeth Hospital, Woodville SA 5011, Australia 1 To whom correspondence should be addressed The optimal time period for intrauterine insemination (IUI) in relation to either luteinizing hormone (LH) surge or human chorionic gonadotrophin (HCG) administration leading to the best pregnancy rates has not been determined. In this study, 856 consecutive human menopausal gonadotrophin (HMG)-stimulated and 49 natural unstimulated IUI cycles carried out at a reproductive medicine unit affiliated with a tertiary centre were analysed in a retrospective fashion. There were three scenarios in the temporal relationship of the LH surge, HCG administration and artificial insemination. These were (group A) subjects who had an endogenous LH surge but were not given HCG; (group B) subjects who were given HCG after an observed LH surge, and (group C) subjects who were given HCG before the LH surge. The overall pregnancy rate (PR) was 16% per cycle. The PR was 9% in group A, 20% in group B and 14% in group C. The PR in group B was significantly better than group C (P 0.04). In group B, the longer the time interval between the LH surge and HCG administration, the better the PR up to 20 h (P 0.025); the timing of IUI based on the LH surge was not critical to the achievement of pregnancy within 3 days. In group C, PR improved with the increasing interval between HCG and IUI from <28 h up to 60 h. We conclude that a better PR is achieved if a spontaneous LH surge occurs before HCG administration, especially where the administration of HCG is delayed 8 20 h after an observed LH surge; the timing of IUI based on the LH surge is not critical to the achievement of pregnancy within 3 days. Key words: HCG/HMG/infertility/intrauterine insemination/ LH surge a higher pregnancy rate (PR) is achieved when IUI is combined with the use of human menopausal gonadotrophin (HMG) to stimulate multiple follicle development (Martinez et al., 1991; Irianni et al., 1993). The efficiency of IUI with HMG injection may be more evident in the treatment of mild or even moderate male factor, cervical factor, endometriosis or unexplained infertility (Chaffkin et al., 1991; Hurst et al., 1992; Nulsen et al., 1993; Gregoriou et al., 1995), but the contributing factors to maximize IUI PR remain unresolved. In this study we examine the influence of some factors on PR. The timing of insemination in IUI programmes in relation to other major events around ovulation or most probably ovulation itself has been suggested as the most important variable affecting the success of this treatment (Allen et al., 1985). To maximize the chance of success, the timing of insemination needs to be closely related to the time of ovulation (Allen et al., 1985; Moghissi, 1986; Kenmamm et al., 1987). There have been some discrepancies in the literature regarding the time of ovulation after the onset of the luteinising hormone (LH) surge. In a multicentred collaborative study from the World Health Organization, it was found that ovulation occurred between 24 and 56 h after the onset of the LH surge (WHO, 1980). Another study found that ovulation occurred at a mean time of 27.3 h from onset of the LH surge (Garcia et al., 1981). Ovum retrieval 36 38 h from the initiation of the LH surge achieves good fertilization rates (Testart et al., 1981). Martinez et al. (1991a,b) suggested that a beneficial effect would arise if the final process of natural follicular maturation was allowed to occur following the spontaneous rise of LH. However, human chorionic gonadotrophin (HCG) is widely used in most IUI programmes to promote the final maturation of pre-ovulatory follicles and to time ovulation. Follicular rupture usually occurs ~36 48 h after HCG injection (Testart, 1990). We aimed to test the hypothesis that a spontaneous LH surge gives a higher PR and that there was an optimal time period in relation to either LH surge or HCG injection during which insemination will achieve the best PR. Introduction Therapeutic intrauterine insemination (IUI) is widely used for the empiric treatment of infertility. Stimulated IUI has been suggested by some authors to give significantly better results compared with timed intercourse (Hurst et al., 1992; Nan et al., 1994). The main hypothesis postulated to explain the success of this treatment has been related to the timing of the presence of an increased concentration of motile spermatozoa in the proximity of the oocyte (Moghissi, 1986). Furthermore, Materials and methods Subject selection This study was based on data collected from subjects treated in the IUI programme in the Reproductive Medicine Unit at the Queen Elizabeth Hospital during 1990 1995. During this period, 463 subjects underwent 856 completed cycles with HMG and HCG to induce ovulation (94%) and 49 natural unstimulated cycles (6%). Usually no more than three cycles were performed per subject. In our Unit, IUI with low dose gonadotrophin stimulation was 2162 European Society for Human Reproduction and Embryology

LH surge and HCG on IUI outcomes mainly indicated for patients with unexplained infertility or moderate semen defect where at least two of the following criteria were generally met in two semen analyses: 15 10 6 spermatozoa/ml, 30 50% progressive motility and 15% normal morphology. All couples had 1 year of primary or secondary infertility and they were investigated by means of cycle tracking, which included a sperm cervical mucus contact test, tubal assessment (either by laparoscopy and dye or hysterosalpingogram) and sperm immobilization test for serum antisperm antibodies. Criteria for inclusion were: (i) female partner should be 45 years of age at the time of treatment and have a normal ovulatory history or ovulate in response to medication; (ii) female partner must have bilateral patent Fallopian tubes, demonstrated within the last 2 years; (iii) male partner must have had at least three semen analyses and one trial sperm washing within the last 2 years. In this study, we located 1090 treatment cycles initiated, including 185 cancelled cycles, not included in this study, and 905 cycles with insemination completed. Of the 905 cycles, 10 cycles (1%) were excluded for missing data and 15 cycles with missing the time of HCG administration were included in the overall analysis, but excluded from some calculations. Sperm preparation Patients were requested not to have intercourse for 3 days before the day of semen collection. Semen samples were produced by masturbation and collected into sterile containers. After complete liquefaction at room temperature, each sample was analysed using WHO guidelines. Semen was prepared for IUI by collecting a motile sperm fraction using either a swim-up procedure or discontinuous Percoll gradients. Samples were prepared on Percoll gradients if the sperm concentration was 20 10 6 /ml and/or progressive motility was 30%. The isolated fraction of motile sperm was prepared in a 0.5 ml aliquot of HEPES-buffered human tubal fluid medium for insemination. Ovarian stimulation and timing of insemination Patients were administered gonadotrophin (HMG, Humegon; Organon, Melbourne, Australia; or Pergonal; Serono, Sydney, Australia) unless they specifically requested a natural cycle. Administration of HMG was started on day 5 of the patient s cycle or the next day if prior measurement of plasma oestrogen and progesterone were low. The routine commencement dose was 75 IU HMG. Blood samples were taken on about day 10 of the patient s cycle to measure oestradiol and LH concentrations. A transvaginal ultrasound scan was performed when serum oestradiol was 0.6 nmol to monitor the number and size of follicles. The dose of HMG was appropriately increased or decreased in response to the oestradiol concentrations in the blood samples and follicle development. If there were three or more follicles with diameter 16 mm on the scan, the treatment was cancelled due to the increased risk of multiple pregnancy. If there was no LH surge by 0900 h (defined in the next section) but 1 or 2 follicles were present with a diameter 16 mm, HCG 5000 units (Profasi: Serono) was given in the evening and IUI was performed 36 48 h after HCG administration. If the LH surge occurred but there was no progesterone shift (defined in the next section), HCG 5000 IU was given at the same day and IUI was performed 30 32 h after the blood test. If the LH surge occurred with a progesterone shift, HCG was administered as soon as possible and IUI was performed within 24 26 h of the blood test. Insemination procedure, luteal support and detecting of pregnancy The cervix was exposed with a bivalve speculum, and a Tomcat catheter (Sherwood Medical, St Louis, MO, USA), which had been curved to the uterine shape, was used to deposit the motile spermatozoa ~0.5 cm below the uterine fundus. Luteal support was provided by administering 1000 IU HCG every 3 days for three doses after the insemination. If no menstrual period occurred, a quantitative β-hcg test was performed 16 days after insemination to determine the establishment of the biochemical pregnancy and clinical pregnancy was confirmed later by ultrasound scanning. In this study, a pregnancy was defined as one with the presence of an embryonic sac confirmed by ultrasound scanning. Collection of data Some standard demographic data were recorded such as age, body mass index [BMI; weight (kg)/height (m) 2 ] and pregnancy. If there was a doubling of LH values and no progesterone rise from baseline, the LH surge was taken as having occurred on the morning of the day blood was taken (i.e. 0800 h). If there was a doubling of the LH values and 1.5-fold increase in progesterone values over the basal value, the LH rise was taken as having occurred 6 h earlier (i.e. 0200 h). If there was a doubling of both the LH and progesterone values over the base value, the LH rise was taken as having occurred 12 h earlier (i.e. 2000 h previous day). All other variables such as date and time of HCG and IUI were calculated from the information recorded in the database. Statistical analysis Statistical analyses were performed using Instat, a statistical program by Graph Pad (San Diego, CA, USA). Pregnancy rate per cycle was compared using the Fisher s exact test or the χ 2 -test, and in all cases P 0.05 was considered significant. Analysis of variance (ANOVA) was used to compare differences of the mean age, BMI and cycles among the groups. Results A total of 463 couples who underwent 905 completed cycles of IUI between 1990 and 1995 was studied. The mean age of the female patients was 31.4 0.3 years. The overall PR was 16% per cycle. Three scenarios in the temporal relationship of the LH surge, HCG administration and artificial insemination were identified, namely: (group A) subjects who had an endogenous LH surge but not given HCG; (group B) subjects who were given HCG after an observed LH surge; and (group C) subjects who were given HCG before the LH surge. The PR of the three groups and other results are presented in Table I. The highest PR was found in those patients in group B (20%), which was significantly higher when compared to group C (P 0.04). The three groups were comparable with respect to age, cycle number and BMI (ANOVA). In group A, 49 cycles (89%) were not stimulated with HMG (natural cycles) and resulted in four pregnancies (8.0%). HMG was used in the remaining 856 cycles (six cycles with HMG only from group A and 850 cycles with HMG and HCG from groups B and C) which resulted in 141 pregnancies (16%). In group B, we calculated the time interval from the onset of the LH surge and the time of HCG injection to the time of IUI. The analysis showed that the timing of IUI based on the LH surge was not critical to PR within 3 days (Figure 1), but Figure 2 indicates that a higher PR was achieved if insemination was done within 48 h of HCG administration. Further analysis of the relationship between the interval between the LH surge and HCG administration and PR is shown in Figure 3. This 2163

K.W.Fuh et al. Table I. The mean ( SEM) age (years), body mass index (BMI, kg/m 2 ), number of cycles (N), mean number of cycles per subject (n) and pregnancy rate per cycle in the three different groups Group Age (years) BMI (kg/m 2 ) N n Pregnancy (%) A: LH surge but no HCG 30.9 5.0 23.8 4.0 55 1.7 1.1 9 B: HCG after LH surge 31.7 5.0 25.3 5.4 403 1.6 0.8 20 C: HCG before LH surge 31.0 4.6 24.8 5.2 437 1.9 0.9 14* *P 0.05 compared with group B. LH luteinizing hormone; HCG human chorionic gonadotrophin. Figure 1. Distribution of the time intervals between onset of the luteinizing hormone (LH) surge to insemination versus pregnancy rate in group B (human chorionic gonadotrophin after LH surge). There was no significant difference on pregnancy rate within the overall range. The number above each bar is the number of cycles in the subgroup. IUI intrauterine insemination. Figure 3. Distribution of time intervals from onset of luteinizing hormone (LH) surge to human chorionic gonadotrophin (HCG) injection versus pregnancy rate in group B (HCG after LH surge). The number above each bar is the number of cycles in the subgroup. Fifteen cycles were excluded due to imprecise recording of time of HCG administration. Figure 2. Distribution of time intervals from human chorionic gonadotrophin (HCG) injection to intrauterine insemination (IUI) versus pregnancy rate in group B (HCG after luteinizing hormone surge). No significant difference was evident between the three groups (P 0.69). The number above each bar is the number of cycles in the subgroup. Figure 4. Distribution of time intervals between human chorionic gonadotrophin (HCG) injection to insemination versus pregnancy rate in group C (HCG before luteinizing hormone surge). The apparent linear trend in this group was not significant. The number above each bar is the number of cycles in the subgroup. One cycle was excluded due to imprecise recording of time of HCG administration. administration and the IUI (Figure 4). There appeared to be a linear trend in this group but it was not significant. shows that the later HCG administration after the LH surge (group B) the higher the PR until ~20 h. There was a significant linear relationship of the 0 8, 8 12, 12 16 and 16 20 h groups to PR (P 0.025). A mean PR of 24% resulted if the administration of HCG was delayed between 8 h and 20 h after the LH surge. In group C, the majority of patients received IUI between 36 and 48 h after elective HCG injection. The lowest PR (5%) occurred at an interval of 28 h between the HCG 2164 Discussion IUI combined with gonadotrophin stimulation has been advocated as an effective treatment for some types of infertility when compared with no stimulation (Martinez et al., 1991b; Chaffkin et al., 1991; Irianni et al., 1993). Whether this improvement in fecundity rates after HMG stimulation was related to the obtained polyovulation only or might have been caused by the correction of subtle ovulatory disorders still remains unclear. In this retrospective study, we looked at

LH surge and HCG on IUI outcomes certain parameters to make sure that they were not affecting the results dramatically (e.g. mean age, mean BMI and mean number of cycles), but certainly we did not and cannot look at every possible confounding factor. Our overall PR was 16% per cycle, which is comparable with some past studies where PR of 14 20 % have been achieved (Chaffkin et al., 1991; Irianni et al., 1993). In our study, subjects who had natural cycle IUI (group A) had requested the procedure either to avoid gonadotrophin stimulation or because they had become pregnant in the past on natural cycle IUI when this was practised in the Unit. They are therefore a select group and probably not comparable to groups B and C. There has been considerable research into the timing of ovulation. Past research has shown that, since there was a wide variation in the time at which the LH peak occurred, the onset of the LH surge was a more suitable reference point in timing ovulation. Testart and Frydman (1982) in a large study found that the time of follicular rupture was almost always subsequent to the 37th hour after the onset of the LH surge. Seibel et al. (1982) reported that human oocytes matured in vitro reached metaphase II after 36 h and in vivo the oocyte may take up to 38 h from the onset of the LH surge before metaphase II was reached. The WHO in a multicentred, collaborative study (WHO, 1980) found that ovulation may occur from 24 to 56 h after the onset of the LH surge, with a mean time of 32 h. Weinberg and Wilcox (1995) estimated that mean viable lifetime in vivo for sperm was 1.4 days, while the lifetime of the ovum appears to be less than a day. Moghissi (1986) found that the fertilizing life span of the ovulated ovum was up to 12 h after ovulation. Much of the variability in timing for ovulation after the LH surge can be explained by the varying definitions of the LH surge and different timing methods. This may suggest that an oocyte can reach maturity and retain the potential to be fertilized over a long fertilizing window (mean 32 h ovulation time plus less than 24 h life span). In our results, we observed the occurrence of spontaneous LH surges in stimulated cycles to be 48%. The LH surge occurred 78% of the time between day 9 and day 13 of the cycle. We analysed the time interval from onset of the LH surge to time of insemination in group B. There was no significant change in pregnancy rates over these time intervals (from 24 to 60 h). This time frame was in agreement with the above-mentioned potential fertilizing window and held true when HCG was given electively. Several studies have stressed that an LH surge is a crucial event for final maturation and ovulation (Tayor et al., 1995; De-Koning, 1995). Martinez et al. (1991a,b) in their controlled study of IUI suggested that the LH surge is a reflection of the natural maturation process of the follicle. They also found that the mean diameter of the pre-ovulatory follicles was significantly smaller when early HCG injections were given compared with cycles with the spontaneous LH surge. According to the general induction criteria, the HCG administration usually took place earlier in the cycles when compared with the occurrence of a spontaneous LH surge. Other data suggest that early administration of HCG may have some detrimental effects (Hillier et al., 1985; Tarin and Pellicer, 1992). Casper et al. (1988) studied a large number of in-vitro fertilization cycles and found higher pregnancy rates where a spontaneous LH surge occurred compared to the HCG-stimulated cycles (28.8 vs 13.9%, P 0.001). In our study, the results supported the above findings: when HCG had been given after the LH surge (i.e. allowing for full maturation of the follicle to occur), a higher PR resulted (group B: 20% vs group C: 14%, P 0.05). Further analysis in that group showed that the timing of insemination based on the LH surge was not critical to the achievement of pregnancy within 3 days, but if it is based on the time of HCG administration a higher PR may be achieved if insemination is performed within 2 days (PR: 20 vs 9%; Figure 2). The other analysis within group B found that the longer the time interval up to 20 h between the LH surge and subsequent HCG administration the better the PR up to 30 38% (P 0.025). In group C, the PR appeared to improve with the increasing interval between HCG and IUI from 28 h up to 60 h (not significant). Although the trend was very clear, the majority of cycles was in the time period of 40 44 h and there were only 20% of the cycles earlier or later than this time. Thus a proper testing of the relationship can only be possible in a prospective trial, although the relationship between LH surge (or HCG administration) and ovulation would suggest that later IUI is more likely to be synchronized with ovulation. It has been our clinical practice to administer HCG to all stimulated cycles even if a spontaneous LH surge occurred. It may be questioned whether this practice is of value and a randomized trial would be justified to prove any benefit of HCG in this situation. It is also debatable as to whether elective administration of HCG for ovulation induction is justified given our results. The real reason for this practice is to enable IUI to occur at a controlled time comparable with clinical and laboratory convenience. Given the large window in which insemination can occur with good PR, there appears to be little justification other than to reduce the amount of endocrine monitoring required to detect the timing of an LH surge. In conclusion, the highest PR was achieved in the group where the LH surge occurred before HCG administration, especially when HCG administration was delayed 8 20 h after an observed LH surge; for those patients with an LH surge, the timing of insemination based on the LH surge was not critical to the achievement of pregnancy within 3 days, but if based on the time of HCG administration, a higher PR was achieved by inseminating within 48 h of HCG administration. Acknowledgements The main author thanks the 807 Military General Hospital, Taipei, Taiwan, Republic of China for providing him generous sponsorship to conduct the work described in this paper. References Allen, N.C., Herbert, C.M., Maxson, W.S. et al. (1985) Intrauterine insemination: a critical review. Fertil. Steril., 44, 569 580. Casper, R.F., Erskine, H.J., Armstrong, D.T. et al. (1988) In vitro fertilization: diurnal and seasonal variation in luteinizing hormone surge onset and pregnancy rates. Fertil. Steril., 49, 644 648. Chaffkin, L.M., Nulsen, J.C., Luciano, A.A. and Metzger, D.A. (1991) A comparative analysis of the cycle fecundity rates associated with combined 2165

K.W.Fuh et al. human menopausal gonadotrophin (hmg) and intrauterine insemination (IUI) versus hmg or IUI alone. Fertil. Steril., 55, 252 257. De-Koning, J. (1995) Gonadotropin surge-inhibiting/attenuating factor governs luteinizing hormone secretion during the ovarian cycle: physiology and pathology. Hum. Reprod., 10, 2854 2861. Garcia, J.E., Jones, G.S. and Wright, G.L. (1981) Prediction of the time of ovulation. Fertil. Steril., 36, 308 315. Gregoriou, O., Vitoratos, N., Papadias, C. et al. (1995) Controlled ovarian hyperstimulation with or without intrauterine insemination for the treatment of unexplained infertility. Int. J. Gynaecol. Obstet., 48, 55 59. Hillier, S.G., Siddiquey, A.K. and Winston, R.M. (1985) Fertilization in vitro of cumulus-enclosed mouse oocytes: effect of timing of the ovulatory HCG injection. Int. J. Fertil., 30, 34 38. Hurst, B.S., Tjaden, B.L., Kimball, A. et al. (1992) Superovulation with or without intrauterine insemination for the treatment of infertility. J. Reprod. Med., 37, 237 241. Irianni, F.M., Ramey, J., Vaintraub, M.T. et al. (1993) Therapeutic intrauterine insemination improves with gonadotrophin ovarian stimulation. Arch. Androl., 31, 55 62. Kenmann, E., Bohrer, M., Shelden, R. et al. (1987) Active ovulation management increases the monthly probability of pregnancy occurrence in ovulatory women who receive intrauterine insemination. Fertil. Steril., 48, 916 920. Martinez, A.R., Bernadus, R.E., Voorhorst, F.J. et al. (1991a) A controlled study of human gonadotrophin induced ovulation versus urinary luteinizing hormone surge for timing of intrauterine insemination. Hum. Reprod., 6, 1247 1251. Martinez, A.R., Bernadus, R.E., Voorhorst, F.J. et al. (1991b) Pregnancy rates after timed intercourse or intrauterine insemination after human menopausal gonadotrophin stimulation of normal ovulatory cycles: a controlled study. Fertil. Steril., 55, 258 265. Moghissi, K.S. (1986) Some reflections on intrauterine insemination. Fertil. Steril., 46, 13 15. Nan, P.M., Cohlen, B.J., te Velde, E.R. et al. (1994) Intrauterine insemination or timed intercourse after ovarian stimulation for male subfertility. A controlled study. Hum. Reprod., 9, 2022 2026. Nulsen, J.C., Walsh, S., Dumez, S. and Metzger, D.A. (1993) A randomized and longitudinal study of human menopausal gonadotrophin with intrauterine insemination in the treatment of infertility. Obstet. Gynecol., 82, 780 786. Seibel, M.M., Smith, D.M., Levesque, L. et al. (1982) The temporal relationship between the luterinizing hormone surge and human oocyte maturation. Am. J. Obstet. Gynecol., 142, 568 572. Tarin, J.J. and Pellicer, A. (1992) Oocyte maturation in human in vitro fertilisation programmes. Ann. Acad. Med. Singapore, 21, 492 497. Tayor, A.E., Whitney, H., Hall, J.E. et al. (1995) Midcycle levels of sex steroids are sufficient to recreate the follicle-stimulating hormone but not the luteinizing hormone midcycle surge: evidence for the contribution of other ovarian factors to the surge in normal women. J. Clin. Endocrinol. Metab., 80, 1541 7 Testart, J. (1990) Artificial insemination. Rev-Prat., 40, 2711 2715. Testart, J. and Frydman, R. (1982) Minimum time lapse between luteinizing hormone surge or human chorionic gonadotrophin administration and follicular rupture. Fertil. Steril., 37, 50 53. Testart, J., Frydman, R., Feinstein, M.C. et al. (1981) Interpretation of plasma luteinizing hormone assay for the collection of mature oocytes from women: definition of a luteinizing hormone surge-initiation rise. Fertil. Steril., 36, 50 54. Weinberg, C.R. and Wilcox A.J. (1995) A model for estimating the potency and survival of human gametes in vivo. Biometrics, 51, 405 412. World Health Organization (1980) Temporal relationships between ovulation and defined changes in the concentrations of plasma oestradiol-17β, luteinizing hormone, follicle-stimulating hormone, and progesterone. Am. J. Obstet. Gynecol., 138, 383 30. Received on April 14, 1997; accepted on July 23, 1997 2166