Contemplating the evolution of the varied clinical uses of pituitary gonadotropic hormones reveals a number of interesting assumptions, and absence of assumptions, that have led us to present-day practices. While strategies for stimulating ovulation have steadily improved over the past 50 years, the basic principles governing clinical practice, e.g., indications for use, protocol design, and precautionary limits and hazards, were almost completely defined during the first decade of use. One exception to this might be the more recent protocol modifications adopted to stimulate ovulation for the purpose of in vitro fertilization (IVF).
To be sure, some protocol modifications represent true progress, namely the use of progressive “step-up” dosing in certain anovulatory situations, and also the recent success with patient self-administration. On the other hand, other modern developments, such as the arrival of a recombinant synthetic product, that importantly diversifies the supply source, have brought little effective change to either the fundamental principles of gonadotropin use or indeed to the success of ovulation stimulation.
Furthermore, some of the allegedly modern advances in protocols can be traced back to earlier concepts that were simply “forgotten” for a time. Examples include the concept of an “FSH threshold,” essentially proposed by Brown in 1978 , the “FSH window,” explored by Lunenfeld as early as 1961 , and the work of Donini testing preparations of varying FSH/LH ratios in 1968 . In addition, the practice of “coasting” therapy as a way to lessen the risk of ovarian hyperstimulation and multiple pregnancies was being used by Delafontaine in 1978 even before the critical advance of ultrasound imaging .
Through the years, gonadotropin preparations for clinical use have been derived from four successive sources; the two most recent remain in use .
1.1.1 Pregnant Mare’s Serum Gonadotropins (PMSG)
Preparations of PMSG first appeared in 1937 and successful clinical use to stimulate ovarian follicular growth in patients was achieved as early as 1941. However, a rapidly-developing immune response was soon recognized, that rendered patients insensitive to further treatment after only a few trials. As a result clinical uses of this product were soon abandoned.
1.1.2 Human Pituitary Gonadotropins (HPG)
Following the clinical failure of PMSG, ovulation stimulation was attempted with an extract of human pituitary tissue. First applied in 1958, the extracts achieved modest success and continued in use until 1988, when the cases of associated Creutzfeldt-Jakob disease became evident. Although clinical use of this preparation would have always been limited by a meager supply of source tissue, these pituitary extracts remain useful today as International Reference Preparations.
1.1.3 Human Menopausal Gonadotropins (HMG)
The ability to isolate and purify hormones from the urine of post-menopausal women ushered in the true era of gonadotropin use. Using a kaolin-acetone extraction method, Bruno Lunenfeld demonstrated as early as 1945 that it was possible to isolate and purify gonadotropins from menopausal urines. Injections of the extract into immature rats stimulated ovarian follicular growth and testicular spermatogenesis in the testis. Although the potential clinical value of these preparations was easy to predict, it remained impossible to patent this natural human product and thus difficult to convince pharmaceutical companies to engage a project that required massive urine collections, construction of an industrial-sized extraction technology, and development of the necessary safety precautions, based solely on hypotheses derived from animal experiments.
The true history of HMG development finally began in Rome in 1957 when Pietro Donini, a senior research scientist who had extracted HMG at the Instituto Farmacologico Serono, invited Professor Lunenfeld to visit and discuss with the institute’s board the feasibility of mass producing HMG and initiating clinical trials. Initially the proposal was not well received, due to little enthusiasm to convert laboratories into commercial urinals. Fortunately, the Vatican had been holding a major share in the Instituto Farmacologico Serono since 1952, and Don Giulio Pacelli was their representative to the board. He also happened to be a nephew of the Pope and began to take an interest in the project. Professor Lunenfeld was asked to remain in Rome for continued discussions that led to a proposal that the homes of retired nuns could provide ample supplies of urine. The Serono board was convinced to undertake the project.
This offer proved to be very successful: by 1961 Lunenfeld was able to report the first newborn infants from ova stimulated by HMG injections , and by this time three collection centers, in Italy, Spain, and the Netherlands, were already operating with more than 600 donors. This same method of HMG production for direct ovarian stimulation continues in use to the present time.
For many years human menopausal gonadotropins remained the only direct ovarian stimulator, and the essentials of stimulation strategy were written through its use. Because the post-menopausal woman is hypogonadal and hypergonadotropic, FSH and LH appear at significant levels in urine. Only a very limited degree of hepatic and renal metabolism occurs, which allows retention of most biological properties.
FSH isoforms in urinary HMG are slightly more acidic, partly because its extremely elevated secretion in hypoestrogenic women contains more of the acidic isoforms, and partly because the acidic isoform population survives hepatic and renal metabolism better. LH content of HMG extracts is proportionately lower than FSH because post-menopausal women secrete more FSH than LH. In order to balance therapeutic concentration of the two hormones closer to equality, it was initially necessary to add pregnancy urine-extracted hCG, the only other hormone preparation available at that time with true LH-like activity. This supplementation was actually unidentified but was suspected for a long time by clinicians. It has only recently been acknowledged by pharmaceutical companies because radioimmunoassay technology has made it possible to distinguish pituitary LH from hCG. hCG supplementation of HMG preparations is no longer permitted and the pharmaceutical manufacturers now claim that a balanced ratio of FSH/LH is achieved by the presence of pituitary hCG that supposedly occurs naturally in post-menopausal urine.
Initially FSH potency varied substantially in the vials supplied by the pioneering manufacturers Serono (Pergonal®, later Neopergonal®) and Organon (Humegon®) that led to variable clinical responses as well. This was likely the result of using a wide variety of urine donors and extraction procedures, and also of using imprecise bioassays to establish international reference units (IU). Vials labeled as containing 75 IU each of FSH and LH could in fact range from 55 to 80 IU of either hormone. Clinicians often blamed these potency variations as the cause of irregular results from IVF procedures, although early embryology technology also brought a number of uncertainties.
Up through the 1980s, the preparations of hMG permitted a satisfactory cure for most dysovulatory or anovulatory patients, save those afflicted with polycystic ovarian disease (PCOD). Initially known as the Stein-Leventhal syndrome, PCOD is typified by excess body weight, anovulation, hyperandrogenism, and an excessive LH secretion that produces a lower serum FSH/LH ratio, along with a possible metabolic syndrome with an enhanced risk of Type II diabetes mellitus. Pathophysiology of this syndrome was believed to result from a relative shortfall of FSH secretion accompanied by LH excess, and it was initially proposed that treatment with pure FSH would restore successful ovulations.
Although this assumption was later shown to be incorrect, it did lend energy to resolving the purity problems of HMG, and a more purified FSH product (uFSH) appeared in 1988. Nevertheless, this preparation offered no therapeutic improvement. It took a return to the original principles of defining an FSH threshold, and the design of a slower “step-up” protocol, to restore normal ovulations in PCOD patients. It turned out that rates of both success and complications were quite the same, whether hMG or uFSH were used .
Development of a refined uFSH preparation became the last example of a progressive step initially sought by clinicians, and to which the manufacturers responded. In a considerable shift of roles, most all of the subsequent developments in gonadotropin preparations have been initiated by the pharmaceutical companies themselves. In the meantime, the market for gonadotropin preparations has been greatly expanded due to increased use for in vitro fertilization procedures. From a “craft” developed and practiced by a relatively small number of gynecologic specialists, the therapeutic approaches to infertility have blossomed into a substantial industry, and include a new set of philosophies and practices. This has resulted in significant conflicts between product design and marketing on the one hand, and uses of therapeutic agents and therapeutic goals on the other.
Coincidentally, marketing strategies put increased pressure on scientific and medical issues. It is important to keep in mind the recent episode of LH slander that occurred simultaneously with the arrival of uFSH. Although pituitary LH was long known as a valuable physiologic promoter of steroidogenesis in ovarian follicles and corpora lutea, the hormone became suddenly and repeatedly attacked during many congresses and symposia as a deleterious therapeutic agent, allegedly responsible for miscarriages, infertility, and other forms of reproductive failure. Although the campaign was disguised as honest scientific debate, it appears to have been little more than a scheme of adverse publicity to discredit hMG preparations. The fact that many clinicians were dragged in good faith into this marketing ploy clearly demonstrates the power that pharmaceutical companies have acquired in order to control initiatives in research and development. Through positions of influence conferred by sponsorship of professional conferences and clinician training programs, corporations continually tout the advantages of their new products. The LH example should cause clinicians to retain an open mind when confronted with these short-lived “truths” and assertions.
Although medical practice welcomed the arrival of purified urinary FSH (uFSH), and later highly purified urinary FSH (uFSH-HP), there remained a problem of supply. Thus the development of genetically synthesized recombinant FSH (rFSH), followed soon by rLH and r-hCG, now offers a valuable and potentially inexhaustible supply of these hormones. However, as typically befalls existing pharmaceuticals, the arrival of “something new” is accompanied by heavy publicity designed primarily to discredit the existing medications they seek to replace. This is regrettable because the urinary gonadotropin extracts were never clinically inadequate, and never deserved to fall from grace. They remain highly valuable FSH preparations.
1.1.4 Recombinant FSH (rFSH)
Once the genes coding for the two subunits of FSH were isolated, it was a simple task to transfect cDNA into appropriate cell lines. Two Chinese hamster ovary (CHO) cell lines, one containing two plasmids and the other one plasmid, are used to produce, respectively, follitropin alpha (Gonal F®) and follitropin beta (Puregon®). As a result, many challenges and problems of natural hormone extracts have been resolved through genetic engineering. The supply is now inexhaustible, batch-to-batch variation has been eliminated, and there is no contamination by foreign (urine) proteins. On the other hand, some of the more recent advances by pharmaceutical companies, namely defining FSH potency by mass rather than bioassay units, or releasing products with specific FSH/LH ratios such as 2/1, or developing longer-acting preparations, all appear to be targeted more toward postponement of patent expirations than to increasing interest among clinicians and their patients. An example of this so-called progress is the filling of FSH injection pens with an identified mass quantity (micrograms) of hormone rather than with international units, in an effort to suppress all batch-to-batch variation that does remain possible with bioassay measures. Such a sophistication is substantially dampened in daily clinical practice by the large fluctuations in ovarian response that occur from one treatment cycle to another in the same patient using the same FSH dosage.
In summary, advances in ovarian stimulation since the appearance of hMG may be summarized in two basic ways: (1) the use of “step-up” low doses and chronic low dose protocols, particularly for PCOD patients; (2) the ability of patients to self-administer the more stable synthetic product, thus permitting infertile couples to participate together in a more personalized therapeutic schedule.
One potential technological advance that has little likelihood of occurring is the availability of LH for triggering ovulation. Unfortunately, the presently-used hCG, whether as a urine extract or a recombinant product, is primarily responsible for ovarian hyperstimulation and multiple pregnancies. Despite a final wish by clinicians to be able to trigger ovulation with synthetic LH or an LH-like molecule, these advanced preparations seem unlikely to become available for that purpose, not because the means to produce it are lacking, but rather because there is little profit incentive to conduct the necessary research. Although manufacturers continue to assert their concern for patient safety, on this particular issue they will continue to do nothing.
1.2 Monitoring Ovarian Stimulation
During the many years when only hMG was available, the principal advance of methodology was the very critical ability to monitor responses to the stimulation: first by measures of follicular estrogen secretion, and then by morphologic criteria using reliable ultrasound imagery.
Initially, estimation of the ovarian response to gonadotropins was possible only through observation of clinically available estrogen target tissues, e.g., vaginal mucosa and the uterine cervix. Examination of target tissue responses such as the vaginal eosinophilic index, cervical mucus, and opening of the cervical external os, were unfortunately of little help, because of typically large variances between patients. Even within the same patient, a dose-response relationship could be imprecise, especially in cases when a maximal response was achieved at relatively low estrogen levels. Significantly improved assessment was not provided even when combined with other clinical observations, such as ovarian volume and sensitivity through vaginal examination. These difficulties explain much of the early occurrence of ovarian hyperstimulation syndrome that unfortunately ended on occasion with a patient’s death.
1.2.1 Monitoring by Hormonal Measures
The first revolutionary advance toward improved monitoring used measures of ovarian estrogen production, initially estimated indirectly by quantifying urinary excretion of estrogen metabolites. Although long available, the usual assay typically required a lengthy turn-around time of up to 3 days. Assay protocols included a prolonged hydrolysis step followed by extraction of the phenolic steroids from urine, but this was of little use to a therapeutic situation where estrogen production could vary continually. Attempts at a more rapid hydrolysis with boiling NaOH proved even less useful because the estrogen analytes were partially degraded, and assay results were tempered.
Ruffié and Jayle deserve considerable credit for improving the steroid hydrolysis step in 1965, by increasing the concentration of a hydrolytic enzyme catalyst (obtained from snails) as well as the temperature of the reaction. Adopting a Kober-Ittrich spectro-fluorometric analysis of the extract yielded a good same-day result from urine specimens collected during the previous 24 h .
Availability of this “rapid” total estrogen assay began to diminish the importance of some imprecise clinical observations, save perhaps the evaluation of cervical mucus, and it produced a much better risk estimate for ovarian hyperstimulation. Because the optimal margin of safety was close to the normal preovulatory estrogen level of 30–100 μg/24 h, it became important to add a creatinine assay to assure that a complete 24 h sample had been collected. By 1968, the sampling interval could be safely shortened to 12 h, and to even as little as an overnight urine collection, which additionally gave greater comfort to a patient’s life. A creatinine measure was still used to adjust for a complete day’s output. Estrogen excretion remains fairly constant throughout the day, and the main gonadotropic effect would be expected within the first 8 h following an evening administration.
Prior to the widespread commercial availability of gonadotropins, only a small number of physicians benefitted from exclusive hormone distributions by the Serono Corporation. These individuals helped start the “Gonadotropin Club” (G-Club) which proposed the term HMG as well as the definitions of the First, and then the Second, International Reference Preparations (IRP) of HMG. Early members of the G Club included A. Netter and R. Palmer (Paris) and I. Bernard (Bordeaux) in France, and G.S. Jones in the USA .
Use of gonadotropins gradually increased through the 1960s albeit still at a limited rate. Statistical evaluation was not yet feasible because the few clinicians who were willing to attempt HMG stimulation were typically unable to present more than a few patients at scientific conferences. Discussions centered primarily on the safety range of estrogen levels, and optimal administration protocols (e.g., whether HMG should be administered daily or on alternate days, and how many vials should be used each time). The continuing major concern was how to provide adequate ovulation while minimizing the risk for hyperstimulation.
Other associated issues, that continue to be debated today, included:
· Whether a “slight” hyperstimulation might enhance the rate of successful pregnancy. It was argued that the range for the 12 h urine estrogen assay could safely be raised to 75–150 μg/24 h, which is significantly above physiologic pre-ovulatory levels.
· The importance of the FSH/LH ratio in HMG. The ratio was typically close to 1 in the usual vials of Pergonal® or Humegon®, but Donini produced a Neopergonal® preparation in which FSH content could be been enriched from 2/1 to as much as 15/1 over LH content. Vials used in France during this time typically contained FSH/LH at 2/1, and provided satisfactory results . However, although an FSH excess was not detrimental to ovarian stimulation, clinicians were reporting that preparations with an FSH/LH ratio of less than 1/1 had a tendency to provoke premature luteinization of the Graafian follicles.
Additional issues of concern included an occasionally severe hyperstimulation syndrome that could occur with estrogen levels well within the safety range, or that multiple pregnancies resulted in some instances where estrogen levels were not different from typical single pregnancies.
The availability of rapid plasma estradiol radioimmunoassay technology in the 1970s shed some light on these questions. Urine excretion of total estrogen is of course delayed in contrast to plasma levels and, although there is good overall correlation during periods of steady hormone secretion, a rapid rise in plasma might not be reflected in urine until the following day (Figs. 1.1 and 1.2). Thus a normal urine value (representing output during the previous night) could be observed while a spike of blood estrogen was occurring from a rapidly developing ovarian hyperstimulation. In response to this, Delafontaine and Grenier proposed a protocol modification designed to reduce risk for hyperstimulation and multiple pregnancy, namely to stop the HMG administration when the plasma estradiol level reached the safety range, and to administer hCG only when estrogen levels actually began to decrease (Fig. 1.3). This became the beginning of the concept of “coasting.” .
Concordance of urinary and plasma estrogen levels (Courtesy of A. Ruffié)
Examples of two patients illustrating the potential for discrepancy between urinary and plasma estrogen assays that may occur when estradiol secretion is rising rapidly (Courtesy A. Ruffié)
A proposed protocol to reduce the risk of ovarian hyperstimulation and multiple pregnancy. Daily HMG injections are halted when plasma E2 has risen sufficiently, and hCG is administered only after the E2 level begins to decline 
Reliable and sensitive assay techniques that used radioiodine tracers after extraction appeared in 1984–1985. A direct assay method using 131I appeared in 2000, followed by the enzyme-linked immunoassays in 2002 that remain in use today. At the present time, measures of plasma estradiol by immunoassay is a simple, rapid procedure that reports a result within a very few hours of sampling, and daily secretion can be monitored, as desired.
1.2.2 Monitoring by Morphology
The final important advance for monitoring was the appearance of ultrasound imagery. Introduced into gynecology practice near the end of the 1970s, this procedure provides direct observation of the number, location, diameter and appearance of the developing ovarian follicle(s) . With this technology it was soon revealed that similar plasma estradiol levels can occur with a variety of follicular situations, for instance, with one dominant follicle and a few smaller ones, or several mid-sized follicles, or with numerous small follicles.
In essence, ultrasound identifies best the risk for multiple pregnancy, while rapid estradiol measures guard best against risk for hyperstimulation . Ultrasound imagery also permits estimation of endometrial grade and thickness, the amount of cervical mucus, and also of the ovarian reserve of antral follicles during in the early days of the cycle. It is still too soon to know whether an automated 3-dimensional count of all developing follicles will improve accuracy and safety even more, compared to the usual manual measurements.
There is no doubt that the combination of functional (plasma estradiol) and morphologic (ultrasound) criteria provides the optimal assurance of both efficacy and safety for an ovarian stimulation. In addition, measures of plasma LH and/or progesterone may help to identify the exact timing of ovulation for cases of intra-uterine insemination, or when only a single intercourse is possible. LH data are also important when monitoring a controlled ovarian hyperstimulation (COH) for IVF purposes with a GnRH antagonist protocol. Some claims have been made that conducting only hormonal or only ultrasound monitoring, or even neither, can suffice for an adequate stimulation leading to IVF, but no feasibility study has successfully challenged the necessity of a using both hormonal and ultrasound for optimal monitoring of ovarian stimulation in the routine practice of reproductive medicine.
Brown JB (1978) Pituitary control of ovarian function. Concepts derived from gonadotropin therapy. Aust N Z J Obstet Gynaecol 18:47–54CrossRef
Lunenfeld B, Rabau E, Rummey G, Winkelsberg G (1961) The responsiveness of the human ovary to gonadotropins (Hypophysis III). Proc Third World Congress Gynecol Obstet (Vienna) 1:220
Donini P (1968) Action de FSH et de LH sur la stimulation folliculaire. Définition du Neo-Pergonal. In: Les Gonadotrophines d’origine humaine. Byla-Searle Ed, Paris
Grenier J, Delafontaine D, Cornier E et al (1981) Régulation de l’induction de l’ovulation par l’association HMG-HCG. In: Hormonologie de la Stérilité – Explorations et Thérapeutique. SEPE Ed, Paris, pp 583–618
Ludwig M, Felberbaum RE, Diedrich K, Lunenfeld B (2002) Ovarian stimulation: from basic science to clinical application. Reprod Biomed Online 5(Suppl 1):73–86PubMedCrossRef
Lunenfeld B, Sulimovici L, Rabau E, Eshkol A (1962) L’induction de l’ovulation dans les aménorrhées hypophysaires par un traitement combiné de gonadotrophines urinaires ménopausiques et de gonadotrophines chorioniques. C R Soc Fr Gyncol 32:346
Sagle MA, Hamilton Fairley D, Kiddy D, Franks S (1991) A comparative randomized study of low-dose human menopausal gonadotropin and follicle stimulating hormone in women with polycystic ovarian syndrome. Fertil Steril 55:56–60PubMed
Dubecq JP, Labadie P, Mathieu H, Ruffie A (1962) Modification apportée à la réaction colorée des phénolstéroïdes urinaire selon Ittrich. Bull Soc Pharm Bordeaux 101:159–188
Bernard I, Ruffie A (1973) Notre expérience des traitements modernes des stérilités par trouble de l’ovulation et les risques d’hyperstimulation. Bord Med 6:2549–2557
Hackloer DJ, Fleming R, Robinson HP et al (1979) Correlation of ultrasonic and endocrinologic assessment of human follicular development. Am J Obstet Gynecol 133:122–128
Emperaire JC, Vella M, Ruffie A (1982) Apport de l’échographie ovarienne dans la surveillance de l’induction de l’ovulation. Ses rapports avec la surveillance hormonale. Rev Franç Endocrinol Clin 23:379–384