If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
I describe how we identified the need to block the luteinizing hormone (LH) surge when trying to control the processes of luteinization and ovulation within the clinic. The first step, in fact, was using ovarian ultrasound evaluation of follicular development in the natural cycle (published in 1979) and then when the ovary was stimulated with exogenous follicle stimulating hormone. We observed that induced multiple follicular development often led to “premature” LH surges—which occurred before the leading follicle had achieved normal preovulatory dimensions. The work required both ovarian ultrasound and reliable radioimmunoassays, which were not always available. When early work with gonadotropin releasing hormone agonists showed that they could suppress LH activity, it was the logical step to try to use them to perform that task during the induction of multiple follicular development. High frequency administration of the gonadotropin releasing hormone-agonist successfully achieved sustained LH suppression through the follicular phase allowing clinical control of luteinization and ovulation.
The year 1977 is a landmark for the understanding of mammalian reproduction because of the Nobel prize for Andrew Schally, with his colleague Roger Guillemin, being awarded for his remarkable endeavors and achievements with gonadotropin releasing hormone (GnRH) and its analogs. However, it is often overlooked that they shared the honor on that day with Rosalyn Yalow, who was awarded it “for the development of radioimmunoassays of peptide hormones.” This latter contribution was key to the establishment of the singular professional sub-specialty of “reproductive endocrinology” simply because her work led to our ability to measure hormones at extremely low concentrations in small amounts of biological fluids. This was a revolutionary step, and it is where my contribution began.
The year before their joint recognition (1976), I moved to Scotland to join a small team working on infertility among women with normal menstrual rhythm to carry out postdoctoral work in a dedicated laboratory associated with the university department of gynecology. The concept (funded by the Scottish government) required me to set up assays and measure steroid and gonadotrophin hormones in women under various conditions. The “dedicated” reproductive endocrinology laboratory was located in a rather unprepossessing part of the Royal Infirmary on the east side of Glasgow. The building itself dated from 1794 and had seen little investment over the following 200 years, except for some beautiful fume cupboards constructed using polished oak bearing some exotic and imaginative graffiti. The team was led by the remarkable Ian Coutts, with whom I enjoyed many years of work and fun: although, despite many attempts, I could never quite match his lifelong commitment to Scotland’s main export.
The Radioimmunoassays and Polyclonal Antibodies
For many years now, immunoassays have deployed monoclonal antibodies with high specificity, often targeted at specific parts of molecules with great precision, and able to operate with high degrees of sensitivity. Although a great step forward, the original radioimmunoassays deployed polyclonal antibodies from various mammalian serum sources, and the characteristics of these amazing tools were complex, with each iteration having unique characteristics, and they could also vary between batches of the same anti-serum. This was particularly important for the larger peptide hormones, including the gonadotrophins. Most luteinizing hormone (LH) assays were characterized or compromised by high degrees of cross-reaction with human chorionic gonadotropin (HCG) and/or free alpha sub-unit of the gonadotropins. The former can be present in high concentrations, so even a low degree of cross-reaction can lead to confusing measurement values.
The alpha sub-unit, common to all gonadotrophins and thyroid stimulating hormone, is rarely present in free form in high concentrations, but there are specific circumstances when they can contribute to confusing observations. Figure 1 shows what happens after starting GnRH-agonist treatment when measured using assays focused on the whole LH, focused mainly on the beta-subunit, compared with one focused on the alpha-subunit alone. It is easy to see how confusing profiles can be experienced when the assay antibody may give undue weight to the alpha sub-unit when measuring what is thought to be LH. This can lead to difficulties in interpretation and undermine confidence in the results.
Figure 1Changes in the concentrations of LH and free Alpha Sub-unit measured using polyclonal antibodies, over 10 days of treatment with Hoe766 (5 × 100 μg per day, intranasally). LH = luteinizing hormone.
The steroid hormones also showed cross-reactions with other steroids of similar size and complexity and progesterone assays with other C21 steroids and corticosteroids, which can be present in significantly higher concentrations. This means that even if there is a low degree of cross-reaction, the effect on the results can be significant. These complexities meant that the interpretation of apparent results could be confusing, and we were always seeking antibodies with properties that suited our tasks better. Ian Coutts was a well-connected individual, and he had a remarkable ability to “find” people who had antibodies that they were willing to provide from many and various sources and also to negotiate “reasonable” terms of purchase: some rather unconventional. It was my role to assess the validity and functionality of the tests for the hormones, or other analytes, we planned to measure.
Normal Cycle Hormone Profiles and Ultrasonic Observation of Follicular Diameter
My work began with a proposal to define normal reproductive hormone profiles throughout the menstrual cycle and explore differences in women suffering from a minimum of 3 years of infertility. These same infertile women were then to be stimulated with modest doses of urinary human menopausal gonadotropins (HMG) in the early follicular phase, with responses monitored by daily serum samples analyzed for the same hormones. Although this strategy includes their own control data, we could not conclude that the data we were to collect represented “normal.”
It was a matter of great fortune that a young German doctor named Jochen Hackeloer contacted us to assert that he thought he could identify follicles growing in the ovary using the ultrasound equipment in Ian Donald’s department on the west side of the city. I have to admit that the picture he showed me looked like a poor quality representation of the lunar surface rather than a follicle! So, we set up a project in which he would measure the size of follicles (follicle diameter [FD]) in a series of volunteers, and we would measure the corresponding estradiol (E2) concentrations. He formed the “ovary club” of numerous remarkable volunteers, who provided his scan data and serum samples: this meant that we had our blood samples from “normal” volunteers, so I was very pleased that he had come to us.
Come the study day of reckoning, when the E2 measurements were correlated with Jochen’s FD estimates, we were profoundly disappointed that there was no correlation between the 2 primary measurements.
However, I had also used the samples to determine normal values of LH and progesterone, using the best antibodies I had tested thus far. The others were unaware that I had these data, so when I suggested that we could examine the follicular phase data in isolation, I was suddenly very popular. Later that day, when we rearranged the data by reference to the LH peak, the data transformed magically into a perfect correlation between the FD and E2. I think of this as a critical moment in my personal history, and the work was soon reported in the world's first demonstration of ultrasound measurement of ovarian follicles published in 1979, causing international interest (
The rearrangement of data meant that the measurements of the corpora lutea and the luteal phase samples were removed from the dataset. Overall, it meant that we could be satisfied that our LH, E2, and progesterone assays were reliable at physiological concentrations, and we suddenly had a new tool by which to explore ovarian function.
Dimensions of the Preovulatory Follicle
This new tool gave us a novel and unexpected aspect to examine, both in questions regarding normality and under stimulation. Critically for future events, we noted that the size of the periovulatory follicle in the normal cycle on the day of the LH surge was between 18 and 26 mm in diameter, with a mean of 22 mm. Correspondingly, we concluded that a FD of 20 mm should be set as a target size range for the lead follicle during ovarian stimulation.
Premature Luteinization
Jochen Hackeloer departed to further his career, but other young researchers were keen to follow up his work with us, as we added FD data to the resources in the study of the effects of early follicular phase HMG treatment in women with normal menstrual rhythm. It meant that we could monitor follicular growth simultaneously with an assessment of estrogen output.
The initial observations were not as simple as we hoped, as the maximum FD seen at the time of the LH surge was often much less than in the normal cycle.
Figure 2 shows an example of the effect of HMG injections on days 1, 3, and 5 on estradiol, LH, and progesterone: the demonstration of premature luteinization. The “premature” adjective derived from the fact that the follicles were so much smaller than normal at the time of the LH surge compared with that in the normal cycle (in this case, 14mm vs.22 mm in the normal cycle).
Figure 2A case demonstrating premature luteinization (the early rise in progesterone) when the increased estradiol concentrations elicited a surge of LH when the lead follicle was only 14 mm in diameter. Normal follicular phase ranges for LH and progesterone are denoted by the shaded areas, as is the normal midcycle peak of estradiol. LH = luteinizing hormone.
We concluded that the premature LH surge was a frequent phenomenon under ovarian stimulation and that, correspondingly, we did not have clinical control of either the ovulatory step or oocyte maturation.
The work was presented in an abstract form in 1979 and 1980 and received little attention at the time. We were not alone in making the observation and conclusion, as Carl Gemzell and colleagues also reported the phenomenon in 1982, having presented evidence in the abstract form before that (
It was clear to us already that if we wanted to have clinical control of the timing of ovulation and egg maturation, we needed a means of blocking the LH surge.
Early Studies with the GnRH Agonist
With perfect timing, Hamish Fraser, in our neighboring capital city of Edinburgh, reported that treatment of macaques with an agonistic analog of GnRH effectively suppressed ovulation through its ability to suppress LH activity. This work was eventually published in 1980 (
Decreased pituitary responsiveness and inhibition of the luteinizing hormone surge and ovulation in the stumptailed monkey (Macaca arctoides) by chronic treatment with an agonist of luteinizing hormone-releasing hormone.
), but his earlier abstracted data had provided the evidence we needed to attempt to use the same GnRH-analog to block the LH surge during ovarian stimulation in women. The work in Sweden with the agonistic analog of GnRH demonstrating its potential role as a contraceptive provided a further indication of its possible role in our setting (
The rights to the use of this drug had been taken up by Hoechst AG, a German chemicals company moving into the world of life sciences. Persuading them that we thought we had specific potential use for their drug was not simple, probably because they thought its potential role was to promote gonadotrophin output rather than exploit its paradoxical down-regulatory effect. However, in the end, they did provide us with some product for initial experiments.
The original product (referred to as “Hoe766”, eventually becoming “Buserelin”) was supplied as a nasal spray, administering 100 μg per spray. The short half-life of Hoe766 meant that we needed a high frequency administration over a protracted period to ensure LH surge suppression, but we did not know how long would be required. We initially explored the effect of 5 × 100 μg and 3 × 100 μg sprays over the waking hours. Figure 3 shows the effect on LH of the flare effect under the 2 regimes and the immediate response to nasal spray administration depending on the day of treatment under the “5-sniff” protocol. After a week of treatment, the absolute LH concentration was within the low-normal range, and the immediate response to the application had become negligible. The “3-sniff” protocol was less reliable, and at no stage did we detect complete elimination of the short-term response.
Figure 3The effects of intranasal administration of the GnRH-analog “Hoe766” on LH over the first week of administration (Panel A) using 2 dosing schedules and the immediate response to administration of one of the applications after increasing duration of treatment (panel B: “5-sniff” protocol). LH = luteinizing hormone.
The “flare” effect of treatment initiation lasted approximately 3 days. It dictated that we should start treatment in the luteal phase so that the flare effect would influence only luteal secretions. Exogenous follicle stimulating hormone (FSH) administration could be started after menstruation, when pituitary responsiveness should be greatly diminished, within a pharmacologically induced hypogonadotropic environment.
Our first use of the GnRH-analog in the luteal phase led to an interesting observation that caused a mild diversion from which I personally learned much. The observation was that the “flare” effect after treatment initiation was luteotrophic: increasing progesterone concentrations—which contrasted with early reports saying that it was luteolytic (
). This caused some consternation at the time among people with higher profiles than our own, and some exchanges between us could be described as “tricky,” almost resentful. Furthermore, when some groups tried to replicate the observations, the measurements of LH frequently showed no concentration decline, implying that their assay was unsuitable for the task and rendering interpretation and mutual agreement difficult.
Demonstration that the LH Surge Could Be Blocked During Ovarian Stimulation
The initial experiments on the effects of GnRH-agonist administration (Fig. 3) were performed in the absence of follicular stimulation, so it remained to be shown that the LH surge could be blocked in the presence of high estrogen concentrations experienced under treatment with exogenous gonadotropins. The knowledge that the response to the agonist could be effectively eliminated within a week of treatment initiation meant that we could proceed to stimulated treatment with urinary-derived gonadotropins (HMG).
Figure 4 shows the events and hormone profiles throμgh one of the first patient’s treatment cycles. The circulating LH remained basal throughout the period of raised estrogen exposure. This meant that, for the first time, we were in the clinical control of the timing of ovulatory events, and we could advise timed intercourse accordingly. Patient feedback at this stage was often entertaining but sadly not recorded. They were clearly stressful times, but their resilience and commitment were extraordinary.
Figure 4Profiles of estradiol, progesterone, and LH through a stimulated treatment cycle showing supranormal concentrations of estradiol and suppression of LH and progesterone until after clinically-induced luteinization induced by HCG. The shaded areas show the normal ranges for the normal cycle. LH = luteinizing hormone, HCG=human chorionic gonadotropin.
We anticipated that the luteal phase output of progesterone after the trigger bolus of HCG might be compromised if the circulating LH were suppressed, so we administered low dose HCG injections 3, 6, and 9 days after the ovulatory bolus. It can be seen from the duration of the raised progesterone output that this degree of luteal support was excessive and added little to the outcome of the cycle except for prolonging it unnecessarily.
Scrutiny of the LH profiles after HCG administration in Figure 4 also shows that the LH values apparently increased in the luteal phase. We contended that this reflected the cross-reaction with the administered HCG within the LH assay. The long half-life of HCG can be seen in these results, and the effect on extending the life of the corpus luteum is well demonstrated in the response of the progesterone. Luteal support was reduced after that.
Efficacy as a Treatment for “Unexplained Fertility”
The initial cohort of 5 patients who volunteered for this protracted process, giving us daily blood samples throughout the procedures, were given a maximum of 3 treatment cycles with timed intercourse. All 5 cases conceived and achieved a live birth (including one set of twins). You can imagine our excitement at this stage. I reported the data at the MRC Unit of Reproductive Biology in Edinburgh when David Baird implored us to publish the data as they stood (
) so that those wanting to pursue “controlled” ovarian stimulation for either ovulation induction or potentially for in vitro fertilization (IVF) could see the potential advantages of the methodology. Of the next cohort of 5 cases, there were 3 further pregnancies, 2 of which were twin pregnancies. The treatment method was extended to women with polycystic ovary syndrome and hypothalamic amenorrhea. The larger series with differing diagnoses were reported in 1988 (
Combined gonadotropin-releasing hormone analog and exogenous gonadotropins for ovulation induction in infertile women: efficacy related to ovarian function assessment.
). Multiple pregnancies, of course, became a burning issue—both in ovulation induction and, eventually, IVF.
The program also moved to undertaking intra-uterine insemination instead of timed intercourse, although it was unclear whether it showed significant clinical advantage at that early stage (
Superovulation with intrauterine insemination in the treatment of infertility: a possible alternative to gamete intrafallopian transfer and in vitro fertilization.
We also reported the first “short course” stimulated cycle, in which ovarian stimulation was initiated on cycle day 3 by starting the GnRH agonist at that stage, followed by hMG stimulation 3 days later. This first case we studied achieved a pregnancy and a live birth, providing us with daily blood samples throughout her conception cycle. However, it became clear in subsequent cases that the flare effect of treatment initiation could resurrect the old corpus luteum and generate high levels of progesterone during the early follicular phase, prohibiting the chance of pregnancy. This effect could be eliminated by before treatment with Norethisterone, but we eventually concluded that it achieved negligible clinical benefit in general use.
The Move To IVF and the Debate Over Ovarian Stimulation
The move to IVF was the objective of many groups at this time, including our own. We were cognizant of the debate regarding the requirement for increased egg numbers while controlling the timing of ovulation and luteinization. The phrase “controlled ovarian stimulation” (COS) had not been invented at this stage, despite our promotional efforts. Furthermore, it seems strange to recall that many senior voices were set against the requirement for multiple mature eggs and the potential benefits of ovarian stimulation.
In the early stages, the 2 primary candidates for ovarian stimulation were Clomiphene, championed by Lopata in Australia (
). However, the impact of spontaneous LH activity during ovarian stimulation when estradiol became supranormal meant that LH monitoring was required, and egg collections frequently had to be performed at very short notice after complex monitoring steps (
Interpretation of plasma luteinizing hormone assay for the collection of mature oocytes from women: definition of a luteinizing hormone surge-initiating rise.
). Often, these monitoring steps were affected using urine samples, which reflect endogenous activity only after a delay of approximately 12 h, so numerous patients were subjected to egg collections that were already too late.
In our own monitoring program with HMG alone, we assessed ovarian and pituitary activity with 2 serum samples per day, and we quickly observed that the spontaneous premature LH surge could be considerably attenuated during ovarian stimulation, as reported in detail by Messinis et al. (
) in 1986. It meant that the surge might be missed with infrequent sampling, as exemplified in the bottom panel of Figure 5. The normal surge lasts up to 3 days, whereas some showed durations of possibly less than 12 hours. Interestingly, the series showed mean LH concentrations within the normal range, demonstrating how significant individual events can be lost within the larger mass of data.
Figure 5Profiles of progesterone and LH during the late follicular phase were monitored by 2 serum samples each day during ovarian stimulation with HMG in the absence (black lines) and presence (red line) of GnRH agonist cotreatment in women presenting “postmature” eggs. The LH is presented as both mean values and (lower panel) as individual cases (unsuppressed cases), revealing the short, attenuated nature of the spontaneous LH surges. The shaded areas represent the normal ranges for the follicular phase of the normal cycle. LH = luteinizing hormone, HMG= human menopausal gonadotropin.
These attenuated LH surges led to modest degrees of luteinization, as represented by the circulating progesterone, using an assay targeted at the low values seen in the follicular phase. Even so, the mean values were not excessive. The rise in progesterone was deemed of insufficient importance to cancel the cycle, so the patients underwent trigger and egg collection as usual. The absence of pregnancy in this series of cycles is probably less convincing evidence than the reports of “postmature eggs,” which was an independent label ascribed by our embryologist. The eggs were characterized by loss of integrity of the corona cells and dispersed granulosa cells, which probably represent biological responses to the attenuated LH surges identified only because of the high frequency blood sampling. The need for a reliable means of blocking the LH surge in IVF programs was clear, and the GnRH-analog method was the obvious choice.
The first report of the combination of our method in IVF was in 1984 from the group at the Middlesex Hospital in London, who performed their work after consultation with us (
). It remains a matter of enduring disappointment that they did not reference or acknowledge our work or personal support in their publication. This was in distinct contrast to the general atmosphere of constructive information sharing that surrounded most groups working toward what was deemed a shared endeavor.
The wide-spread adoption of GnRH-agonist COS for IVF was rapid thereafter—simply because it allowed clinical control of the processes of luteinization and ovulation. Egg collections could be undertaken on schedule when the lead follicle was of appropriate dimensions, without the anxiety of preoperative ovulation. Monitoring of responses to stimulation could be vastly reduced, and the field was opened for all the major advances that followed, including ultrasound-guided egg collections and improved laboratory culture facilities.
The publication I enjoyed most was that by Jean-Rene Zorn in 1987 in Paris, in which he demonstrated that the sacred French Sunday could remain unsullied by having to perform egg collections because of spontaneous LH activity (
). He very kindly invited me to a couple of meetings in that wonderful city.
The drug companies quickly prepared different formulations and delivery systems, making COS more convenient for patients and efficient within a clinical environment. In 2000, the GnRH antagonists were introduced, allowing further refinement to the treatment programs, as their use required no before treatment and were needed only when there was a risk of the unwanted LH surge.
It remains my personal position that the apparently simple step of blocking the unwanted LH surge is the most important single step on the road to the success story that is IVF today. The elimination of the main complicating factor in multiple follicular development has prepared the ground for remarkable improvements in laboratory conditions and culture media through comparatively uncomplicated scientific analyses.
From the reproductive endocrinologist’s viewpoint, the deployment of GnRH analogs has allowed us to stimulate multiple follicular activity within the constraints of the metabolism of exogenously administered FSH. Despite the stabilized gonadotrophin environment, the long half-life of FSH (circa 30h) means that we have only modest means of controlling the degree of ovarian response to stimulation, as the 24 hour exposure of the ovaries to FSH remains difficult to manage. The introduction of anti-Müllerian hormone as the indicator of response potential is perhaps the last component required to provide the information for clinicians and patients in deciding the best way forward with ovarian stimulation.
Acknowledgments
The author feels fortunate to have worked with many outstanding people over the years and would like to thank them for providing a wonderful work environment. Furthermore, he wants to voice his admiration and gratitude to the volunteers and initial patients who gave so much of their time and commitment, and blood to provide investigators with the information needed. Their efforts are rarely sufficiently recognized, but they can be more than proud of what their commitment led to. Many of the original women are grandparents now and tend to forget how challenging their efforts were, although the author is secure in the knowledge that they are also very grateful.
Decreased pituitary responsiveness and inhibition of the luteinizing hormone surge and ovulation in the stumptailed monkey (Macaca arctoides) by chronic treatment with an agonist of luteinizing hormone-releasing hormone.
Combined gonadotropin-releasing hormone analog and exogenous gonadotropins for ovulation induction in infertile women: efficacy related to ovarian function assessment.
Superovulation with intrauterine insemination in the treatment of infertility: a possible alternative to gamete intrafallopian transfer and in vitro fertilization.
Interpretation of plasma luteinizing hormone assay for the collection of mature oocytes from women: definition of a luteinizing hormone surge-initiating rise.
00118-0&id=gr1.jpg){kind=link}
00118-0&id=gr2.jpg){kind=link}
00118-0&id=gr3.jpg){kind=link}
00118-0&id=gr4.jpg){kind=link}
00118-0&id=gr5.jpg){kind=link}