Recent Progress in Hormone Research: Proceedings of the 1969 Laurentian Hormone Conference

By E. B. Astwood

Recent Progress in Hormone Research, Volume 26 covers the proceedings of the 27th annual meeting of the Laurentian Hormone Conference held at Mont Tremblant, Quebec, Canada on August 24-29, 1969. The book presents papers on the pituitary and gonadal hormones in women during spontaneous and induced ovulatory cycles; the mechanisms regulating the menstrual cycle in women; some physical and hydrodynamic properties of human FSH and LH; and the biological significance of the prostaglandins. The text also includes papers on thyroid peroxidase and thyroxine biosynthesis; the significance of circulating triiodothyronine; the multiple hormone interactions in the development of mammary gland in vitro; and the intranuclear metabolism of testosterone in the accessory organs of reproduction. Papers on the aspects of androgen-dependent events as studied by antiandrogens; the hormonal control of hepatic gluconeogenesis; and the control of fat cell development and lipid content are also encompassed. The book further encompasses papers on the properties of thymosin; the study of spermatogenesis and steroid metabolism in cultures of mammalian testes; and the role of ACTH on the metabolism of adrenal cell cultures.

• Language: English
• Publisher: Academic Press
• Release date: Oct 22, 2013
• ISBN: 9781483219479

Book preview

26–28.

Pituitary and Gonadal Hormones in Women during Spontaneous and Induced Ovulatory Cycles¹

G.T. ROSS, C.M. CARGILLE, M.B. LIPSETT, P.L. RAYFORD, J.R. MARSHALL², C.A. STROTT and D. RODBARD,     Endocrinology Branch, National Cancer Institute, Bethesda, Maryland

Publisher Summary

This chapter discusses pituitary and gonadal hormones in women during spontaneous and induced ovulatory cycles. In a study described in the chapter, a series of spontaneous cycles were selected on the basis of two presumptive indicators for ovulation: (1) a biphasic basal body temperature curve and (2) a postovulatory interval of 13 days or more measured from the day of the LH peak until the onset of the following menses. From daily determinations of plasma concentrations of follicle-stimulating hormone (FSH), luteinizing hormone (LH), 17-hydroxyprogesterone, and progesterone, the chapter describes mean and modal patterns of changes for these hormones. The chapter presents the results of similar studies made during cycles in which ovulations were induced with clomiphene citrate to define the features that were common to spontaneous and induced ovulations. It discusses hormonal events in some cycles during which characteristic features failed to occur.

I Introduction

The president of the World Bank, Robert S. McNamara, recently observed:³ The greatest single obstacle to the economic and social advancement of the majority of the people in the underdeveloped world is their rampant population growth. In appealing for appropriate solutions, Mr. McNamara asked: Are we to solve this problem by famine? Are we to solve it by riot, by insurrection, by the violence that desperately starving men can be driven to? Are we to solve it by wars of expansion and aggression? Or are we to solve it rationally and humanely in accord with man’s dignity?

In responding to the questions raised by Mr. McNamara, my colleagues and I have the conviction that appropriate solutions will be based upon insights gained by scientific investigation of the physiology of reproduction. We know that Dr. Pincus shared our opinion (Pincus, 1965), and we are privileged to present for the second Gregory Pincus Memorial Lecture our studies of hormonal events during the human menstrual cycle. On this occasion we recall that Dr. Pincus was one of the first and foremost in applying the scientific method to the study of reproductive physiology and in the search for solutions to problems of overpopulation: rational and humane solutions, in accord with man’s dignity.

Prior to the development of radioimmunoassays for gonadotropins (Bagshawe et al., 1966; Midgley and Jaffe, 1966; Odell et al., 1966), relatively insensitive methods of measurement coupled with low concentrations of these substances in blood and urine of normal men and women made relatively large quantities of urine or plasma requisite for accurate measurement (Apostolakis, 1960; Becker and Albert, 1965; Igarashi et al., 1967; Keller, 1966; Kulin et al., 1968; McArthur et al., 1964; Yokota et al., 1965). Consequently specimens had to be pooled either vertically within days between subjects or horizontally within subjects between days, so that changes in hormone concentrations occurring at short intervals were obscured. Alternatively, the large numbers of samples from many subjects required in order to detect subtle changes made the undertaking logistically formidable.

Similarly, more sensitive and practical methods for measuring gonadal steroid hormones in plasma were essential, and these methods have been developed (Baird and Guevara, 1969; Kirschner et al., 1965; Korenman, 1968; Murphy, 1964, 1967; Neill et al., 1967; Riondel et al., 1965; Strott and Lipsett, 1968; Yoshimi and Lipsett, 1968). Thus, it is now possible for the first time to consider interactions of components of the hypothalamic-hypophyseal-gonadal axis in individual men and women studied at intervals of 24 hours or less. Development of some of these methods and studies using them have been the principal research interest of investigators in the Endocrinology Branch, National Cancer Institute, for almost a decade (Bardin et al., 1967; Cargille et al., 1968b; T. Davis et al., 1965; Jacobson et al., 1968a, b).

For this study, a series of spontaneous cycles were selected on the basis of two presumptive indicators for ovulation: (1) a biphasic basal body temperature curve (Hartman, 1962), and (2) a postovulatory interval of 13 days or more (Rock, 1949; Vande Wiele and Turksoy, 1965) measured from the day of the LH peak until the onset of the following menses. From daily determinations of plasma concentrations of follicle-stimulating hormone (FSH), luteinizing hormone (LH), 17-hydroxyprogesterone, and progesterone, we have described mean and modal patterns of changes for these hormones. Results of similar studies made during cycles in which ovulations were induced with clomiphene citrate were examined in an attempt to define the features that were common to spontaneous and induced ovulations. We then considered the hormonal events in some cycles during which characteristic features failed to occur and attempted to interpret the phenomena in terms of a hypothesis concerning interactions of the hypothalamic-hypophyseal-ovarian axis in normal women. Discrepancies and similarities between our data and the data others have obtained by measuring hormonal activities in plasma and urine have been described. Explanations have been advanced to account for discrepancies where these were found.

II Materials and Methods

A SUBJECTS

The subjects studied were mostly young women, ranging in age from 18 to 35 years, admitted either as normal controls recruited under the aegis of the Normal Volunteer Patient Program, Clinical Center, National Institutes of Health, or as patients referred for evaluation of hypogonadism, menstrual abnormalities, or infertility. In addition, a few women, some of whom were postmenopausal, admitted either for follow-up or treatment of malignancies, agreed to participate in studies such as measurement of metabolic clearance rates. Most of the normal volunteers were unmarried and most of the patients were married. Fully informed consent was obtained for every procedure undertaken and the safety and scientific relevance of all procedures were reviewed in accordance with standard practice at the Clinical Center, National Institutes of Health.

Basal body temperatures (BBT) were measured either orally or rectally (using thermometers specially designed for such measurements) and recorded daily. No restrictions were placed upon physical activities or diet of either normal volunteers or patients.

Culdoscopy was performed by the method of Marshall and Hammond (1966), for visualization and biopsy of the ovary when these procedures were indicated for diagnostic evaluation.

Endometrial biopsies were examined histologically and dated according to the criteria of Noyes et al. (1950).

For attempts at induction of ovulation in women with oligoovulatory or anovulatory infertility, clomiphene citrate was given orally in divided doses ranging from 25 to 250 mg per day for 5 days (Jacobson et al., 1968a,b).

B ASSAYS

Venous blood was usually taken in the fasting state before 9:00 AM, anticoagulated with heparin, and centrifuged; plasma was stored at −15°C until assayed. Aliquots of the same samples were used for all measurements, and all samples from a single study period in a given subject were measured in the same assay to take advantage of increased within assay precision. In the case of steroid hormones, pooling of aliquots of several daily specimens was sometimes necessary in order to obtain adequate samples.

Plasma FSH and LH were measured by radioimmunoassays (Cargille et al., 1968a; Odell et al., 1966, 1967a), and progesterone and 17-hydroxyprogesterone by competitive protein binding assays (Strott and Lipsett, 1968; Yoshimi and Lipsett, 1968). Results of all determinations of FSH and LH were expressed in terms of units of biological activity of the Second International Reference Preparation for Human Menopausal Gonadotropin (IRP 2 HMG). This material was shown to have the requisite criteria of similarity for use as a reference preparation in assays of plasma samples using our reagents for both FSH and LH. Dose response relationships obtained from assays of varying doses of solutions of this reference preparation dissolved in buffer were the basis for graphic interpolations for potencies of unknown samples.

For LH, a minimal detectable quantity varied from 0.75 to 1.9 milli International Units (mIU) per assay tube, more commonly the latter. In terms of a maximal volume of 300 μl for plasma samples, the lower limits of detection usually ranged from 6 to 8 mIU/ml. The standard deviation of within assay precision averages 1.5 mIU for values ranging from 15 to 40 mIU/ml. The corresponding value for between assay precision is 4 mIU (Rodbard et al., 1968).

For FSH, the minimal detectable quantity varied from 0.57 to 1.93 mIU per assay tube. A 200-μl volume for plasma samples usually permitted detection of at least 4.7 mIU/ml. Based upon duplicate measurements of six plasma samples in 24 assays, within assay precision was ±0.9 mIU/ml and between assay precision was ±2.0 mIU/ml for the range of 5–11 mIU/ml (Rodbard et al., 1968).

The measurements of progesterone and 17-hydroxyprogesterone have a coefficient of variation of 8% at a plasma concentration greater than 30 ng/100 ml. The smallest detectable concentration is usually 15 ng/ml.

C DATA ANALYSIS

The onset of menses has been used as a reference point for beginning and ending sampling periods despite the fact that, in common with Brewer, we believe menses to mark the end rather than the beginning of a menstrual cycle (Brewer and Jones, 1947). However, since many clinically useful data have been obtained using this traditional external marker, it is not easily abandoned.

While onset of menses has been used to delineate sampling periods, another event has been used as a reference point for pooling data, for calculating means and for defining intervals: namely, the day on which the maximal plasma LH concentration is observed. We have elected to use it as a marker since we believe it to be the most accurate indicator of the phases of the ovulatory cycle.

Hereafter, this maximal LH concentration will be referred to as the LH peak, and the day on which it occurs will be designated as day zero in graphic displays of the data. In Fig. 1 this format is shown together with definitions of intervals used. Unless otherwise specified, numbers on the ordinate represent milli International Units (mIU) of IRP 2 HMG per milliliter of plasma for FSH and LH, or nanograms per milliliter of plasma for 17-hydroxyprogesterone and progesterone.

FIG. 1 Schema for synchronizing cycles, calculating means, and determining intervals.

For convenience, we have designated the period from the day following onset of flow until 1 day prior to the LH peak as the follicular or preovulatory phase of the cycle and the interval from 1 day following the LH peak until and including the day of the onset of next menses as the luteal or postovulatory phase. This formulation may be too restrictive for the follicular phase, particularly if this period be reckoned to terminate with rupture of the follicle and extrusion of an ovum since these two events are thought to occur rarely, if ever, prior to a midcycle gonadotropin surge. However, the designation for luteal phase is probably accurate since a distinct rise in progesterone, thought to indicate early function of the corpus luteum, will be shown to occur regularly by the day following the LH peak in each of these cycles.

III Results

A METABOLIC CLEARANCE RATES

Metabolic clearance rates (MCR) for FSH (Coble et al., 1969) and LH (Kohler et al., 1968) were similar for pre- and postmenopausal women whether determined by continuous infusion of labeled hormone to constant specific activity or by the single injection technique (Table I). The MCR for LH is greater than that for FSH when the two are measured either separately as shown in Fig. 2, or simultaneously in a single subject as shown in Fig. 3. Neither MCR is altered by large differences in concentration of the hormone so that variations in plasma concentrations of FSH and LH must be proportional to changes in pituitary secretion. In the data to be presented subsequently, changes in plasma concentrations will be considered as equivalent to changes in pituitary secretion.

TABLE I

Metabolic Clearance Rates for Plasma FSH and LH in Pre-and Postmenopausal Women

FIG. 2 Average metabolic clearance rates (MCR) for FSH and LH determined by a method based upon continuous infusion and by a method based upon a single injection of isotopically labeled hormone.

FIG. 3 Percent of injected dose of FSH-¹²⁵I and LH-¹³¹I remaining in plasma at varying times following simultaneous injection of the two preparations in the same subject. Reprinted from Coble et al. (1969).

B TIME AND FREQUENCY OF SAMPLING

In some cycles studied it was not possible to obtain every sample at the same time of the day. We considered the possibility that marked diurnal variation might introduce another variable into the results obtained in these cycles. Accordingly, concentrations of FSH and LH were measured in samples obtained at 8:00 AM, 4:00 PM, and 10:00 PM daily throughout a spontaneous presumptively ovulatory cycle in each of three young women. One of these cycles is shown in Figs. 4 and 5.

FIG. 4 ) are shown for comparison.

FIG. 5 ) are shown for comparison.

No consistent pattern of variation was seen in either FSH or LH concentrations which was related to time of day. Values for samples taken at any given time appeared to vary randomly in relation to the mean of all values for the day. The changes in pattern were equally apparent in the curves generateci when the 8:00 AM, 4:00 PM, and 10:00 PM values were plotted separately. We conclude that variations in time of sampling introduced no significant bias in these data for either FSH or LH.

The extent to which areas under portions of the curve around the midcycle peak would vary in relation to frequency of sampling was examined. Detailed analysis indicated that the area under the plasma concentration curves was not systematically biased by frequency of sampling.

C STUDIES OF SPONTANEOUS PRESUMPTIVELY OVULATORY CYCLES

1 Gonadotropins

Determinations of hormone concentrations measured during 71 cycles were available for analysis. Inclusion in this portion of the study was based on five criteria:

1. No clinical or laboratory evidence of endocrine disease or gross menstrual irregularity.

2. No drugs known to inhibit ovulation given either immediately prior to, or during the menstrual cycle studied.

3. Continuous daily sampling starting not more than 4 days after the onset of menses that marked the beginning, and continuing until not less than 4 days prior to onset of menses that marked the end, of the sampling period.

4. An interval of not less than 21 days nor greater than 35 days between first day of menses at the beginning, and first day of menses at the end, of the sampling period.

5. Measurement of FSH and LH made in daily samples in assays in good quality control and measurement of plasma progesterone made either in daily samples or in pooled samples taken prior to and for several days after the LH peak.

Of 71 cycles studied prior to June 15, 1969, only twenty-five met these criteria. For inclusion in the group used to define mean and modal patterns of change in hormone concentrations in spontaneous presumptively ovulatory cycles there were two additional requirements:

1. A biphasic basal body temperature curve.

2. An interval equal to or greater than 13 days from the day of the LH peak to onset of the next menses.

The second of these two additional criteria excluded nine cycles, leaving 16 cycles from 15 women for definition of mean and modal patterns of variation in spontaneous presumptively ovulatory cycles. These cycles ranged from 26 to 34 days in length and were distributed around a mean of 29 days (Fig. 6).

FIG. 6 Frequency of cycles of varying length among 16 presumptively ovulatory cycles.

For this group of 16 presumptively ovulatory cycles, mean daily values for LH and the 95% confidence limits of the means are shown in Fig. 7. Two features characterize the LH pattern. First, there is a peak more than twice the mean of values observed prior to or after it. The mean values the day before and the day after the peak are clearly greater than means on all other days before and after the peak. Second, there is a tendency for mean LH concentrations to rise during the follicular phase prior to and to decline during the luteal phase after the peak.

FIG. 7 Mean (bold line) and 95% confidence limits of mean (shaded area) daily plasma LH concentrations in 16 presumptively ovulatory cycles synchronized around the day of the LH midcycle peak.

Daily mean plasma FSH concentrations show a distinct pattern (Fig. 8). Early in the follicular phase, mean values progressively increase, a feature we have called the early follicular rise (Cargille et al., 1968c). This is followed by a decline during the second half of the follicular phase which reaches a low point we have called the preovulatory nadir. The preovulatory nadir is followed by an abrupt rise, the midcycle peak, usually occurring on the day of the LH peak. During the luteal phase mean values gradually decline to the lowest level observed during the cycle, the luteal nadir. In fact for FSH, a progressive decline, interrupted only by the mid-cycle peak, began in the second half of the follicular phase and extended to the luteal nadir. After the luteal nadir a tendency to rise was observed around the time of onset of menses, a feature we believed to represent the onset of the early follicular rise of the ensuing cycle.

FIG. 8 Mean (bold line) and 95% confidence limits of mean (shaded area) daily plasma FSH concentrations in 16 presumptively ovulatory cycles synchronized around the day of the LH midcycle peak.

Figure 9 shows values for FSH and LH plotted together. During the second half of the follicular phase, mean LH concentrations showed a significant upward trend while FSH levels showed a definite downward trend. These changes occurred during the period of rapid follicular growth (Abarbanel, 1957) when plasma concentrations of 17-hydroxyprogesterone (Strott et al., 1970) and estrogen (Abraham, 1969; Brown, 1960; Brown and Matthew, 1962; Korenman et al., 1969) were increasing rapidly prior to the LH peak.

FIG. 9 Mean (bold lines) daily plasma LH and FSH concentrations in 16 presumptively ovulatory cycles synchronized around the day of LH peak. Shaded areas represent 95% confidence limits of the means.

The possibility existed that some of the features noted for both LH and FSH were artifacts of the method of considering the data relative to the LH peak. We therefore examined follicular phase features, using (as a reference point) the day of onset of flow marking the beginning of the sampling period. Similarly, luteal phase features were related to the day of onset of menses marking the end of the sampling period. Only minor quantitative changes in pattern of FSH and LH levels resulted from changing the method of pooling (Fig. 10). Thus, a progressive follicular phase rise in LH concentrations was more clearly evident as was the rise in FSH levels beginning late in the luteal phase. Progressive luteal phase declines in both FSH and LH concentrations were clearly seen when the data were centered on the day of onset of menses marking the end of sampling period.

FIG. 10 Mean plasma LH (upper panels) and FSH (lower panels) concentrations during first 12 days and last 12 days of 16 presumptively ovulatory cycles synchronized around onset of menses at beginning and end of sampling periods. Shaded areas represent 95% confidence intervals of means.

The progressive follicular rise in means of LH concentrations is significant at p < 0.01 using the sequence sign test of Bross (1960). The progressive luteal decline in LH and FSH concentrations and the early follicular rise in both hormones were also significant at p < 0.05. It was thought that the decline in luteal phase concentrations of FSH and LH might be more clearly related to the early follicular phase of the ensuing cycle than to the follicular phase of the same cycle. Accordingly, we began sampling in the luteal phase of one cycle and continued into the early follicular phase of the next cycle in 13 women. In Fig. 11 these values were pooled vertically in relation to the onset of menses (day zero). Here, the early follicular rise in FSH concentration was seen to begin in the luteal phase of the antecedent cycle. A less consistent but clearly evident rising trend in LH values was also seen. Failure to consistently detect a late luteal rise marking the beginning of the rise for a new cycle when patterns of individual cycles were examined, may have been related to variability in the relationship between initiation of the rise and onset of the menses. Thus, our data indicated that while onset of menses clearly delineates termination of an endometrial cycle, use of the menses as a marker for initiation of an ovulatory cycle may be misleading.

FIG. 11 Means of plasma LH and FSH concentrations (bold lines) and their 95% confidence limits (shaded areas) in samples taken daily for 8 days prior to and 8 days after the onset of menses in 13 cycles.

We have become satisfied that these several features observed in FSH and LH patterns represent biologic phenomena rather than artifacts. We have established criteria to be used in our laboratory to estimate the frequency of occurrence of each of these features of the mean gonadotropin patterns in individual cycles (Table II).

TABLE II

Laboratory Criteria for Occurrence in Individual Cycles of Features of Mean FSH and LH Patterns

Changes have been made in the criteria we published for FSH (Cargille et al., 1968c). These older criteria were found too stringent to describe the true frequency of occurrence of features in presumptively ovulatory cycles using our methods of measurement. In addition, we now describe criteria for judging whether all the features of the LH pattern are present in an individual cycle.

Using these criteria, Tables III and IV summarize the frequency with which these features were observed in the individual cycles comprising this group. Variations in number of cycles examined related to the fact that these criteria could not be applied to an occasional cycle because of failure to have a determination for FSH on the day of the LH peak.

TABLE III

Frequency of Occurrence of Features of Variation in Plasma LH Concentrations in 16 Presumptively Ovulatory Cycles

TABLE IV

Frequency of Occurrence of Features of Variation in Plasma FSH Concentrations in 16 Presumptively Ovulatory Cycles

Features common to all cycles were an LH peak, an early follicular rise, and a luteal nadir for FSH. In the majority of cycles, there were a progressive rise and a luteal decline in LH concentrations, and a midcycle FSH peak, and preovulatory nadir in FSH concentrations.

2 Steroids

Having characterized changes in patterns of pituitary hormone concentrations in plasma during these presumptively ovulatory cycles, we turned to a consideration of patterns of variation in plasma concentrations of some gonadal steroid hormones. Results obtained from daily measurement of plasma progesterone in 10 subjects with a post-peak interval of 13 days or more were averaged using the day of the LH peak as the reference point (Fig. 12). Starting from measurable though low values in the follicular phase, mean concentrations began to rise coincident with the LH peak, increased gradually thereafter to a peak on day +8 and finally declined progressively toward, but not quite to, concentrations seen early in the follicular phase.

FIG. 12 Mean (bold line) and 95% confidence limits of mean (shaded area) daily plasma progesterone concentrations during 10 presumptively ovulatory cycles with intervals of 13 days or more from the LH peak until onset of menses. Cycles were synchronized around the day of the LH peak.

Plasma progesterone concentrations were not measured in every sample from the 16 presumptively ovulatory cycles. For some of the cycles, only measurements of concentrations in pools made from samples on days 5 to 9 before and days 5 to 7 after the LH peak were available. Thus, in order to apply a common standard for comparison of values in all cycles, means were calculated for concentrations on days 5, 6, 7, 8, and 9 before and days 5, 6, and 7 after the LH peaks in cycles where these were measured daily. In each cycle, a significant increase in progesterone concentration was observed when values before and after the LH peak were compared. In 15 of 16 instances, the pooled values on days 5 to 7 exceeded 5 ng/ml. In the 16th cycle, plasma progesterone increased somewhat more gradually so that a value of 5 ng/ml was not found until day 8 after the peak when the concentration was 13.3 ng/ml. Thus, a plasma progesterone concentration in excess of 5 ng/ml was seen in every cycle 5–8 days after the LH peak.

Although plasma progesterone concentrations began to rise slowly on the day of the LH peak, the mean increase in progesterone levels occurred when plasma FSH and LH concentrations were falling, the apex of the progesterone curve coinciding with the FSH luteal nadir. This temporal relationship demonstrated that increasing biosynthesis of progesterone was not dependent upon increasing pituitary secretion of FSH and LH during the luteal phase and suggested the possibility that the luteal phase decline in both FSH and LH may have resulted from increasing ovarian steroid secretion.

Increasing plasma concentrations of progesterone were expected to be reflected in increasing basal body temperatures (M. E. Davis and Fugo, 1948; Israel and Schneller, 1950). This supposition was confirmed by examination of Fig. 13 in which means of daily basal body temperatures were plotted with mean plasma progesterone concentrations. Although mean temperatures rose progressively, the increase was not significant in relation to preovulatory means until 4 days after the LH peak. The mean plasma progesterone concentrations at that time were around 2.5 ng/ml, suggesting that concentrations in excess of this value may always result in an elevation of BBT. Mean basal body temperatures remained elevated with respect to mean follicular phase temperatures for 9 days.

FIG. 13 Mean (bold line) daily basal body temperatures (BBT) and mean (bold line) daily plasma progesterone concentrations during 16 presumptively ovulatory cycles synchronized around the day of the LH midcycle peak. Shaded areas represent 95% confidence limits of means.

Significant increments in plasma progesterone concentrations after the LH peak were further presumptive evidence for occurrence of ovulation. Increased urinary excretion of pregnanediol, thought to be derived largely from progesterone, has been regarded as the hallmark of an ovulatory cycle and adequacy of corpus luteum function since the classical study of Vennning and Browne in 1937.

While development and function of a corpus luteum is usually linked with ovulation, defined as maturation followed by rupture of a follicle with extrusion of an ovum, these two processes may be dissociated. This would be difficult to detect if judgment as to occurrence of an ovulation depends either on corpus luteum function and progesterone secretion or on gonadotropin patterns. Hence, a direct index of follicular maturation would be useful for the purpose of examining the processes separately. Patterns of change in FSH and LH concentrations prior to and including the preovulatory peak may be necessary for and consistent with ovulation, but do not provide conclusive evidence confirming that event.

We have reported previously (Strott et al., 1970) that plasma 17-hydroxyprogesterone, secreted by the ovary, increases significantly prior to and coincident with the LH peak during presumptively ovulatory cycles. Further, we have postulated that this increment may be regarded as a measure of follicular maturation (Strott et al., 1970). To characterize the pattern of change in 17-hydroxyprogesterone concentrations more completely, this steroid was measured either daily throughout a cycle or in either a limited number of individual samples or in pools on appropriate days prior to and after the LH peak. Results of such studies in eight women from the series of 16 with presumptively ovulatory cycles are shown in Fig. 14. Mean values began to rise during the second half of the follicular phase, reached a peak coincident with the LH peak, declined for several days immediately after the peak, and rose again before falling precipitously just prior to the onset of menses. This pattern of variation is similar to that observed by Fotherby, who measured urinary pregnanetriol thought to be derived from 17-hydroxyprogesterone (Fotherby, 1962). The curve of plasma 17-hydroxyprogesterone concentrations is similar to that for plasma estrogens (Abraham, 1969; Baird and Guevara, 1969; Korenman et al., 1969) as well.

FIG. 14 Mean daily plasma 17-hydroxyprogesterone concentrations (solid line) during 8, and mean daily plasma progesterone (dashed line) concentrations during 10, presumptively ovulatory cycles synchronized around the day of the LH midcycle peak. Area under progesterone curve is shaded for contrast.

In Fig. 15, the curves of mean 17-hydroxyprogesterone, progesterone, FSH, and LH concentrations and BBT are shown, all centered on the day of the LH peak. Detailed examination of this composite figure is desirable in order that temporally coincident features can be appreciated.

FIG. 15 Mean daily basal body temperatures (BBT) (upper), daily plasma 17-hydroxyprogesterone and progesterone concentrations (middle), and mean daily plasma LH and FSH concentrations (lower) during 16 presumptively ovulatory cycles synchronized aound the day of the LH midcycle peak.

Hormonal changes during spontaneous ovulatory cycles appeared to begin with an increase in plasma concentrations of FSH and LH late in the luteal phase, prior to onset of menses marking the end of the preceding menstrual cycle. After the initial rise, and coincident with an increase in plasma 17-hydroxyprogesterone concentrations, FSH levels declined while LH levels continued to increase progressively. Mean plasma 17-hydroxyprogesterone concentrations began to rise prior to, and reached a peak coincident with, the LH and FSH peaks but declined transiently thereafter.

Coincident with the LH peak, plasma progesterone concentrations began to rise. This increment was associated with initiation of the basal body temperature increase, which reached a plateau prior to maximal elevation in plasma progesterone concentrations. The temporary decline in plasma 17-hydroxyprogesterone concentrations occurring immediately after the LH peak was followed by a secondary rise which was essentially coincident with maximal plasma progesterone concentrations. After the LH peak, elevated values for 17-hydroxyprogesterone and progesterone were coincident with a continuing downward trend in FSH and LH mean values. The nadir for mean FSH concentrations appeared to correspond to the fastigium of mean progesterone concentrations. Finally, the early rise appeared to begin during the decline in 17-hydroxyprogesterone and progesterone concentrations near the end of the luteal phase.

Thus, to the features described for gonadotropins, we now add features descriptive of variations in plasma 17-hydroxyprogesterone and progesterone concentrations during presumptively ovulatory cycles.

1. For 17-hydroxyprogesterone

a. Rising concentrations during the second half of the follicular phase when LH concentrations are increasing reaching a peak coincident with the LH peak followed by

b. A precipitous decline, which is followed in turn by

c. A secondary rise coincident with maximal mean progesterone levels, occurring despite declining FSH and LH concentrations.

These patterns of variation we have described for concentrations of 17-hydroxyprogesterone in plasma are similar to those others have described for plasma estrogens during presumptively ovulatory cycles (Abraham, 1969; Baird and Guevara, 1969; Korenman et al., 1969).

2. For progesterone

a. An increase beginning coincident with the LH peak when samples are taken at 24-hour intervals and continuing despite progressive decrements in mean plasma FSH and LH to reach

b. A fastigium over the period of 6–8 days after the LH peak coincident with the time of maximal depression in mean FSH concentrations and with a continuing decrease in mean LH concentrations followed by

c. A progressive decline toward follicular phase levels when plasma FSH and LH concentrations begin to rise.

D INDUCTION OF OVULATION BY CLOMIPHENE CITRATE

Since ovulation is cyclic in normal women after puberty, there must be elements which are essential for continuity as well as elements which are essential for a single cycle. Further, within a given cycle there must be elements whose role is primarily the temporal coordination of events in order to maximize the likelihood that fertilization and implantation will follow an appropriately timed coitus during an ovulatory cycle.

For the study of all these components, hormonal events occurring during artificial induction of ovulation are useful for several reasons. First, these usually are studied in women in whom ovulation does not occur regularly, so that events essential to continuity might be identified among them. Second, for some of them, an agent such as clomiphene citrate is sufficient for induction of an ovulation. Thus, occurrence of an ovulation after clomiphene citrate may be interpreted to indicate a functionally responsive hypothalamic-hypophyseal-ovarian axis. Studies of patterns of variation in hormonal concentrations during control and treatment periods should highlight the events that initiate an ovulatory cycle. Finally, studies of patterns of variation in hormone concentrations associated with failures of ovulation induction among these might be expected to provide insight as to the locus of failure and in addition, help to define component interactions necessary for the process both within and between cycles.

Nine apparently ovulatory cycles induced with clomiphene citrate among 34 women with infertility due to anovulation or oligoovulation fulfilled the criteria used to select spontaneous presumptively ovulatory cycles. These were biphasic BBT, interval from LH peak to menses of 13 days or more, measurements of FSH and LH prior to and during the induced cycles, and measurements of plasma progesterone before and after the LH peak. Early in our experience we observed that, as a group, patients with anovulation or oligoovulation had higher mean LH concentrations than our control subjects. However, in common with patterns from spontaneous cycles in normal women, mean postovulatory LH values were significantly lower than mean preovulatory values.

These early impressions have been validated with a wider experience. In addition, marked variability was observed when results of individual studies were compared. Hence, in order to normalize the results for pooling, we have expressed daily values for each woman as a function of percent change from control values obtained in samples taken from the same subject prior to beginning clomiphene citrate.

These numbers have been pooled vertically in the following fashion: For averaging of values for days 1 through 10, all cycles were aligned on the first day of clomiphene administration, day 1. For averaging of values for the luteal phase, all cycles were aligned on the day of the midcycle peak. Both follicular nadir and midcycle peak gonadotropin values and respective days of occurrence were averaged independently and plotted as though occurring on the respective mean days.

Mean relative FSH and LH concentrations rose during clomiphene treatment but declined prior to the preovulatory peaks (Fig. 16). During the postovulatory period, mean concentrations of both hormones declined to a luteal nadir.

FIG. 16 Changes in mean daily plasma LH (dashed line) and FSH (solid line) concentrations during presumptively ovulatory cycles induced with clomiphene in 9 women with oligoovulatory infertility. Values on ordinate indicate percent of control values for LH (left ordinate) and FSH (right ordinate). Days 1–10 were synchronized on the day when treatment began, but subsequent days were synchronized on the day of the LH midcycle peak.

Plasma progesterone levels in either single or pooled samples taken prior to and after the LH peak showed a mean increase comparable to that observed in spontaneous cycles.

Five patients had endometrial biopsies taken after the LH peak, and all showed secretory transformation, judged to be temporally appropriate for the cycle, and in one of them a corpus luteum was biopsied during culdoscopy. When 17-hydroxyprogesterone concentrations were measured in daily samples from three of these cycles, results were similar to those obtained in studies of spontaneous cycles. In the aggregate then, patterns of variation in plasma pituitary and gonadal hormone concentrations during clomiphene-induced ovulatory cycles were similar to those seen in spontaneous cycles with comparable postovulatory intervals. However, several differences merit some consideration.

First, the LH rise beginning during clomiphene administration was not progressive as it was during the follicular phase of spontaneous cycles but declined prior to the preovulatory LH peak. However, it must be recalled that this was only in relation to control values, which tended to be high among these women, and mean actual LH concentrations during the second half of the follicular phase were comparable in the two groups.

Second, mean early follicular concentrations of FSH and LH were higher relative to the midcycle FSH peak or the LH peak than mean concentrations during similar periods of spontaneous cycles. It is tempting to speculate that this phenomenon might be related to either superovulation or ovarian enlargement sometimes noted after clomiphene (Kistner, 1965).

Such a speculation is supported by the third difference, a tendency of values for progesterone and 17-hydroxyprogesterone noted during the luteal phase of patients ovulating following clomiphene to be higher than those observed in spontaneous cycles.

Finally, a rising trend for gonadotropin concentrations near the end of the luteal phase was conspicuously absent, and its absence may have been physiologically significant since sequential ovulatory cycles do not usually follow either single or multiple clomiphene-induced cycles (Kistner, 1965).

Study of these cycles emphasizes the importance of an early follicular increase of gonadotropins in initiation of an ovulatory cycle since no cycle was seen in which an ovulation occurred without a relative increase in plasma FSH and LH concentrations. This observation led us to studies of cycles in which attempts at ovulation induction with clomiphene citrate apparently were unsuccessful.

One such cycle is shown in Fig. 17. There was no increase in either FSH or LH in response to clomiphene, and no significant change in either hormone was observed during the entire period of sampling.

FIG. 17 Daily plasma LH (dashed line) and FSH (solid line) concentrations taken during a cycle during which clomiphene failed to induce an ovulation. Note absence of increase in LH and FSH during clomiphene treatment.

These observations are consistent with the hypothesis that an increase in plasma FSH and LH concentrations is required to initiate an ovulatory cycle. However, an ovulation did not necessarily follow in every instance after the increase in FSH and LH occurred, as Fig. 18 shows. In this cycle, an early rise in both FSH and LH occurred, but they apparently elicited no gonadal steroid response since 17-hydroxyprogesterone concentrations remained unchanged. The absence of a later secondary increase in FSH and LH concentrations suggested that gonadal response to the early rise may be important in linking the early rise to the preovulatory surge.

FIG. 18 Plasma 17-hydroxyprogesterone (solid line) concentrations (in upper panel) and plasma LH (dashed line) and FSH (solid line) concentrations (in lower panel) associated with failure to induce an ovulation with clomiphene. Note that during treatment both LH and FSH concentrations rose, but no subsequent increase in either 17-hydroxyprogesterone, LH, or FSH was seen.

E SPONTANEOUS AND INDUCED CYCLES WITH SHORT LUTEAL PHASES

Selection of cycles for inclusion in the pool for definition of events in spontaneous presumptively ovulatory cycles was based upon (1) An interval of 13 days or more from the LH peak until onset of next flow, (2) biphasic basal body temperature curves, and (3) measurements of plasma progesterone before and after the peak. A group consisting of nine cycles with an interval from the LH peak to menses of less than 13 days were excluded from consideration. Two of these cycles—one with an interval of 11 days and another with an interval of 12 days after the peak—met the criteria for most of the features occurring more than 80% of the time in the group with longer intervals, including appropriate increments in progesterone and 17-hydroxyprogesterone. By these criteria, an interval of at least 11 days after the LH peak would be compatible with occurrence of patterns of variation in hormonal concentrations which we would associate with spontaneous presumptively ovulatory cycles and normal corpus luteum function.

In the seven remaining cycles, the interval after the LH peak ranged from 6 to 8 days and a number of differences became apparent in comparison with the group with longer luteal phases (Table V). The mean duration of these cycles was significantly shorter than that of cycles with longer intervals after the LH peak. Since the mean length of the follicular phase was not significantly different between the two groups, the differences in length were due principally to shorter luteal phases.

TABLE V

Comparison of Lengths of Cycle, Follicular Phases, and Luteal Phases in Normal and Short Luteal-Phase Cycles

ap < 0.001.

Mean plasma FSH and LH concentrations during these short cycles were calculated and plotted in the usual fashion (Fig. 19). Most of the features of the normal cycle patterns were noted for FSH and LH. For LH, these included a progressive follicular rise, a peak, and a luteal decline; and for FSH an early rise, a preovulatory nadir, a midcycle peak, and a luteal nadir. However, two differences were apparent: (1) The FSH midcycle peak followed the LH peak, the maximal mean value being observed on the day following rather than the day of the LH peak. This was due to failure of the two peaks to coincide in most of these cycles. (2) Although luteal phase concentrations of FSH were lower than those in the second half of the follicular phase, there was no tendency for mean FSH values to rise at the end of the luteal phase.

FIG. 19 Mean (bold line) and 95% confidence limits of mean (shaded areas) daily plasma concentrations of LH and FSH during 7 cycles with post LH peak intervals of 8 days or less (short luteal phase cycles). All cycles are synchronized on day of LH peak.

Results of applying the criteria used for the features characteristic of FSH and LH concentrations during presumptively ovulatory cycles to each of these cycles with shorter luteal phases are shown in Tables VI and VII. The FSH midcycle peak coincided with the LH peak in two cycles, and it occurred on the following day in three of five cycles. In one subject, no midcycle FSH peak was noted, and in another the FSH concentration was not measured in the specimen collected on the day of the LH peak. Otherwise, incidence of features was similar in the two groups.

TABLE VI

Incidence of Features of Plasma FSH Pattern in Short Luteal Phase Cycles

TABLE VII

Incidence of Features of Plasma LH Pattern in Short Luteal Phase Cycles

When mean values for FSH for the two groups were compared (Fig. 20), consistently lower mean FSH concentrations were noted for the short luteal phase cycles. Mean LH concentrations tended to be higher prior to the peak but lower on the day of the peak and during the luteal phase for the short luteal phase cycles (Fig. 20).

FIG. 20 Mean daily plasma LH and FSH concentrations (solid line, small points) and their 95% confidence intervals for 16 cycles with post LH peak intervals of 13 days or more (normal luteal phase cycles) compared to mean daily LH and FSH concentrations (dotted lines, solid triangles) for 7 cycles with post LH peak intervals of 8 days or less (short luteal phase cycles). All cycles are synchronized around the day of the LH midcycle peak.

Mean FSH:LH ratios in cycles with long and short luteal phases were compared (Fig. 21). On 22 of 24 days, mean FSH:LH ratios were lower in the cycles with short luteal phases. Although mean LH concentrations were not consistently higher in cycles with short luteal phases, consistently lower mean FSH concentrations served to reduce the mean FSH:LH ratios.

FIG. 21 Mean daily plasma FSH:LH ratios for 16 cycles with post LH peak intervals of 13 days or more, normal luteal phase cycles (solid line, small points), and 7 cycles with post LH peak intervals of 8 days or less, short luteal phase cycles (dotted line, solid triangles). All cycles are synchronized on the day of the LH peak.

In view of the significant difference in FSH concentrations and consequent changes in FSH:LH ratios observed, some differences in ovarian steroid hormone secretion were expected. Mean 17-hydroxyprogesterone concentrations in samples from six of these seven short luteal phase cycles were compared to corresponding values in long luteal phase cycles (Fig. 22). Four features were apparent:

FIG. 22 Mean daily plasma 17-hydroxyprogesterone concentrations during 8 cycles with post LH peak intervals of 13 days or more, normal luteal phase cycles (small points, solid line), compared with corresponding values for 6 cycles with post LH peak intervals of 8 days or less, short luteal phase cycles (small points, dotted lines) synchronized around the day of the LH midcycle peak. Area under the curve for short cycles is shaded for contrast.

1. From the beginning of the rise prior to the LH peak, mean plasma 17-hydroxyprogesterone concentrations for the short luteal phase cycles were lower than corresponding mean values for the normal cycles.

2. The beginning of the rise one or two days prior to the LH peak, seen in the group of longer luteal phase cycles, was not observed in the short luteal phase group. This failure to rise before the peak in the short luteal phase cycles occurred despite increasing mean second half follicular phase plasma LH concentrations equal to, if not greater than, those noted at this time in long luteal phase cycles. This suggested that low 17-hydroxyprogesterone concentrations at this time may represent inadequate ovarian steroidogenic response to gonadotropins in short luteal phase cycles.

3. Maximal concentrations, occurring on the day of the LH peak, were followed by a decrease on the day after the peak in both groups.

4. A sustained secondary rise in 17-hydroxyprogesterone coincident with increasing progesterone concentrations in long luteal phase cycles did not occur in short luteal phase cycles.

Decreased progesterone secretion was also found in short luteal phase cycles. Values observed in samples from five of these seven subjects were compared with mean values in normal luteal phase cycles in Table VIII and Fig. 23. Maximal progesterone concentrations occurred earlier in relation to the LH peak in short luteal phase cycles so that means of concentrations (Fig. 23) on days 5–7 clearly discriminated between the two groups (Table VIII). Further, in none of the short luteal phase cycles did plasma progesterone levels reach 5 ng/ml, the minimal concentration defining cycles with normal luteal phases.

TABLE VIII

Comparison of Mean Plasma Progesterone Concentrations in Pooled Plasma Samples Taken at Varying Times before and after the LH Midcycle Peak during Cycles with Normal and Short Luteal Phases

FIG. 23 Plasma progesterone concentrations in four cycles with post LH peak intervals of 8 days or less, short luteal phase cycles (small points and lines), compared to 95% confidence limits of mean daily plasma progesterone concentrations in 10 cycles with post LH peak intervals of 13 days or more, normal luteal phase cycles (shaded area). All cycles are synchronized on the day of the LH peak.

Although plasma progesterone concentrations were lower, the BBT curves were biphasic in the short luteal phase cycles and sustained during this short luteal phase (Fig. 24). Whether these quantities of progesterone were adequate to regularly produce a secretory transformation in the endometrium is unknown since endometrial biopsies were proscribed in our normal volunteers. In view of the observations of Hertig and his associates (Hertig, 1964; R. F. White et al., 1957) concerning the time of implantation in relation to ovulation and conception, it seems unlikely that implantation would occur regularly prior to menses in such cycles.

FIG. 24 Mean (bold line) and 95% confidence limits of mean (shaded area) daily basal body temperatures (BBT) in 7 cycles with post LH peak intervals of 8 days or less (short luteal phase cycles). Results of frequent determinations of plasma progesterone concentrations in four of these cycles are also shown. Note that all values are less than 5 ng/ml. All cycles have been synchronized on the day of the LH midcycle peak.

One important question raised by these observations related to whether an ovulation occurred during each of these cycles with shorter luteal phases. We would suggest that this happened in some of them since (1) the plasma gonadotropin pattern was almost normal; (2) plasma 17-hydroxyprogesterone increased at the time of the LH surge; (3) some luteinizing occurred, albeit minimal; and (4) in one woman in whom ovulation was induced by clomiphene citrate, a corpus hemorrhagicum was found in association with a typical short luteal phase (Fig. 25). In this cycle, an early rise in FSH and LH was succeeeded by a decline, followed by coincident peaks, and subsequent low levels until the beginning of menses 8 days after the LH peak. An endometrial biopsy taken on the second day after the LH peak showed a secretory pattern consistent with day 16 or 17 of a spontaneous cycle. Plasma 17-hydroxyprogesterone increased coincident with the LH peak, but attempted measurement of plasma progesterone was not successful. Culdoscopy done 3 days after menses began revealed a corpus hemorrhagicum which was biopsied and thought, on the basis of histologic examination, to be approximately 1 week old.

FIG. 25 Concentrations of LH, FSH, and 17-hydroxyprogesterone in plasma samples taken during a clomiphene-induced cycle with a short luteal phase (interval of 7 days after the LH peak). Note that endometrial biopsy on the third day after the LH peak showed secretory change, and that a corpus hemorrhagicum was seen and biopsied on the ninth day after the LH peak.

Even culdoscopic visualization and biopsy of a corpus hemorrhagicum is not unequivocal evidence of ovulation. These observations in a single individual do not constitute evidence that ovulation occurred during all cycles with short luteal phases. However, they are consistent with the view that ovulation occurred during some of them and that ovulation and formation of a corpus luteum either have different hormonal requirements or do not always occur together.

IV Hypothesis Concerning Pituitary–Ovarian Interactions

On the basis of changes in plasma concentrations of pituitary and gonadal hormones observed during spontaneous and induced cycles, we propose the following hypothesis for the interactions of hypophysis and ovary in normal women during their reproductive years. While the two are usually correlated, ovulation may occur without subsequent formation of a corpus luteum and secretion of progesterone so that gonadotropin requirements for these processes may not be identical. For ovulation, spontaneous or clomiphene-in-duced, the minimal pituitary hormonal requirements would be an initial increase in both FSH and LH followed by an appropriately timed pulse of LH. However, an initial increase in FSH and LH does not inevitably result in ovulation since appropriate ovarian response either may fail to occur or may not be followed by a preovulatory surge of LH.

Initiation of the early rise in plasma FSH and LH in spontaneous cycles is related to the steroid hormonal milieu of the luteal phase of the antecedent cycle. Maximal luteal phase concentrations of gonadal steroid hormones such as progesterone, 17-hydroxyprogesterone, and estrogen coincide with minimal FSH concentrations and with a continuing decline in LH concentrations and act in suppressing pituitary hormone secretion. When the concentrations of these steroid hormones begin to decline, pituitary secretion of FSH and LH is resumed and another cycle of follicular maturation is initiated.

In response to increasing pituitary secretion of FSH and LH during the first half of the follicular phase, follicular maturation continues so that by the second half of the follicular phase both a measurable increase in plasma steroid hormone concentrations and follicular enlargement are observed. The steroid hormonal milieu of the second half of the follicular phase suppresses pituitary secretion of FSH but either (1) fails to suppress or (2) stimulates secretion of LH, or (3) reflects the stimulatory effects of increased LH secretion on steroidogenesis.

When steroid hormone concentrations fail to increase during the second half of the follicular phase in clomiphene-induced cycles, no spontaneous midcycle surge is observed. These increments in plasma gonadal steroid concentrations are important to the initiation of a midcycle surge of FSH and LH and constitute a link between follicular maturation and the preovulatory surge. Our data suggest that progesterone is not the steroid responsible for the surge. This linkage may be important in the prevention of hyperstimulation or superovulation in spontaneous cycles. In addition to a regulatory role in ovarian enlargement, the steroid hormonal milieu of the second half of the follicular phase may be important to the role of the LH:FSH surge in follicular rupture.

If the follicle has been adequately prepared by gonadotropin stimulation, and possibly only if an appropriate steroid hormonal milieu is present, the preovulatory surge of LH results in follicular rupture and luteal transformation of both granulosa and theca cells. During the luteal phase, increasing steroidogenesis is associated with either a progressive decline or no change in gonadotropin concentrations. The steroid hormones secreted by the corpus luteum exert their effects on the endometrium as well as at the hypothalamic-pituitary level. Since follicles in the human ovary remain responsive to gonadotropins at all times even during pregnancy (C. A. White and Bradbury, 1965), random follicular maturation and presumably random ovulation could occur in the absence of a mechanism for cyclic suppression of pituitary gonadotropin secretion.

The significance of the mid-cycle FSH rise, commonly but not always seen in spontaneous cycles, is unknown. It usually coincides temporally with the LH peak, and failure to coincide is more common in abnormal cycles. In women with hypogonadotropic hypogonadism, induction of ovulation does not require a midcycle injection of FSH. On the basis of Lostroh’s observations of the efficacy of FSH in stimulating rupture of mature follicles in the hypophysectomized rat (Lostroh and Johnsen, 1966), it is tempting to speculate that coincidence of FSH and LH peaks is a fail-safe redundant feature to assure that rupture follows follicular maturation.

To summarize the essential hormonal requirements for ovulation coupled with formation and function of a corpus luteum: an adequate amount of FSH and LH, perhaps an appropriate ratio of the two, is essential for proper maturation of a follicle. Once adequate maturation has occurred, a surge of LH, or administration of HCG as an LH surrogate, is followed by rupture of the follicle, extrusion of an ovum, and formation of a corpus luteum. Function of the corpus luteum adequate to produce normal secretory transformation of the endometrium does not depend on further increase in pituitary secretion of either FSH or LH in either spontaneous or induced cycles.

V Review of Other Studies

Review of results obtained by others in similar studies seems relevant since contention concerning validity of the model proposed centers in part around apparent discrepancies between our data and other data obtained, both by radioimmunoassay and by bioassay for gonadotropins, and by other methods of measuring ovarian steroid hormones in plasma.

A OF GONADAL STEROID HORMONES

Considering results of measurements of steroid hormones first, the patterns of variation in plasma progesterone concentrations we report are consistent with measurements others have made either of plasma progesterone (Neill et al., 1967; Riondel et al., 1965; Runnebaum et al., 1965; Saxena et al., 1968) or, alternatively, of urinary pregnanediol, commonly regarded as a metabolite uniquely derived from progesterone (Brown, 1960; Brown and Matthew, 1962; Klopper, 1957). Also the patterns of variation in plasma 17-hydroxyprogesterone we report are consistent with Fotherby’s report (1962) of variation in urinary pregnanetriol, thought to be a metabolite of 17-hydroxyprogesterone secreted by the ovary during the normal menstrual cycle. Since patterns of change observed using different methods of measurement of gonadal steroids in plasma are in excellent agreement, protein binding or radioimmunoassays are more practical because of the smaller volumes of plasma required and relative ease of performance.

B OF GONADOTROPINS

1 Total Gonadotropins

In contrast to the consistency of results obtained by various methods for measuring ovarian steroid hormones, results of different methods for measurement of gonadotropins are somewhat more variable and require consideration to ascertain sources of differences. A midcycle increase in total gonadotropin activity in normal ovulatory cycles has been commonly observed despite the variety of biological fluids tested and the variety of methods used for (1) preparing concentrates, (2) pooling samples prior to assay, (3) assaying for activity, and (4) pooling results obtained (Borth et al., 1957; Buchholz, 1957; D’Amour, 1943; Farris, 1946; Johnsen, 1959; Lloyd et al., 1949; Loraine and Bell, 1963; Pedersen-Bjergaard and Pedersen-Bjergaard, 1948; Rosemberg, 1960). In fact, while this increase is not always seen in every cycle studied for reasons that are predominantly methodological in origin, most reports contain some evidence, direct or indirect, for a midcycle rise. In contrast, results of specific assays for FSH and LH by biological and immunological methods have not always been so consistent in relation to the presence or absence of a midcycle increase.

2 FSH

Considering results of immunoassay for FSH, with the exception of failure to show either an early FSH rise or a luteal FSH decline in a report from one laboratory (Saxena et al., 1968), patterns described for changes in plasma concentrations of FSH by radioimmunoassay from seven different laboratories using seven different antisera have been consistent (Cargille et al., 1969; Faiman and Ryan, 1967; Franchimont, 1966; Midgley and Jaffe, 1968; Odell et al., 1968; Taymor, 1968). In the laboratory reporting exceptional results, the anti-FSH antiserum and the method for separating antibody bound and free tracer hormone were unique, and both may contribute to the failure to find features of the pattern observed by others. Patterns similar to those obtained by radioimmunoassays of plasma have been shown in results of an