|Home | About | Journals | Submit | Contact Us | Français|
To characterize the growth and regression of the corpus luteum (CL) during an interovulatory interval (IOI) using serial transvaginal ultrasonography.
Fifty healthy women of reproductive age with a history of regular menstrual cycles underwent daily transvaginal ultrasonography for one IOI. Measurements of luteal area and luteal numerical pixel value (NPV) were recorded each day after ovulation until the CL could no longer be detected. Blood was drawn every third day during the IOI to measure serum concentrations of progesterone and estradiol-17β.
Corpora lutea were of two morphological types: those with a central fluid-filled cavity (CFFC) (78%) and those without (22%). Eighty-eight percent of women exhibited a CL containing a CFFC 2 days after ovulation, followed by 34% 13 days after ovulation and 2% 27 days after ovulation. Luteal area, progesterone concentration and estradiol concentration increased for approximately the first 6 days following ovulation followed by a subsequent decline. Luteal NPV decreased from days 1 to 11 and increased during days 11–16. Changes in luteal area, NPV, progesterone and estradiol concentrations did not differ in women with two versus three waves of follicular development.
Peak luteal function, as determined by maximum luteal area, progesterone concentration and estradiol concentration, is observed 6 days following ovulation. Luteal NPV is reflective of morphological and endocrinological changes in the CL. The development of a CFFC during luteinization is a normal physiological phenomenon. The CL can be detected, but is not functional, during the follicular phase of the menstrual cycle.
The corpus luteum (CL) is a dynamic endocrine gland within the ovary that plays an integral role in regulation of the menstrual cycle and early pregnancy. The CL forms from cells of the ovarian follicle wall during ovulation. The precise origin of the cells that comprise the CL remains controversial. However, it is believed that luteal cells originate from the theca and granulosa cells following breakdown of the basal lamina immediately prior to follicle rupture1.
The human CL was first studied in the 1950s using histological evaluation of ovaries following hysterectomy and salpingo-oophorectomy2. Ultrasonography has since become an invaluable tool for evaluating the CL in vivo. It has been reported that the CL grows to a maximum diameter of 25–40 mm during the luteal phase, as determined by transabdominal ultrasonography3,4. Some CL were reported to exhibit a ‘cystic cavity’ while others did not3–5. Histological evaluation of the CL revealed that there was fresh bleeding into the central cavity following follicle rupture2. The blood-filled CL is referred to as the ‘corpus hemorrhagicum’. Growth of the CL is associated with an increase in luteal blood flow6–8 and serum progesterone concentrations5,8. Peak luteal vascularity appears to occur 7 days after ovulation, the approximate day of maximal progesterone secretion8.
Ultrasonographic detection of the CL following ovulation has been reported to occur in only 50–80% of natural menstrual cycles3,5. With improvements in imaging technology, the ability to visualize the ovulatory process and the immediate site of ovulation has become more feasible9,10. Limited information on the serial characterization of luteal morphology and endocrinological function during spontaneous menstrual cycles using high-resolution transvaginal ultrasonography and endocrine profiling is available. Increased knowledge of the CL during natural menstrual cycles will increase our understanding of luteal dysfunction in women experiencing infertility and/or recurrent miscarriage. New investigations of luteal form and function during the menstrual cycle are further warranted by the recent finding that women exhibit two or three waves of ovarian follicular development during the interovulatory interval (IOI)11,12. An IOI is defined as the interval from one ovulation to the subsequent ovulation11,12. Ovarian follicular waves and luteal morphology in women are comparable to those previously described in domestic animals (i.e. bovine and equine models)13–18. As a result, the bovine and equine species have been developed as models for studying ovarian function in women19,20.
During the bovine estrous cycle, regression of anovulatory follicular waves prior to the ovulatory wave was shown to occur through negative feedback effects of luteal progesterone on luteinizing hormone (LH) pulse frequency and estradiol production21,22. Periodic development of anovulatory follicle waves continued during the bovine estrous cycle until the CL regressed23. Luteal regression occurred later in animals with three versus two follicle waves23,24 and progesterone levels remained elevated longer in three versus two wave animals25. During the equine estrous cycle, plasma progesterone concentrations did not differ between fluid-filled and non fluid-filled CL17. In addition, ultrasonographic image attributes of bovine and equine CL have been evaluated in detail using gray-scale image analysis. Mean numerical pixel value (NPV) of the CL decreased during the growth phase26,27 and increased during luteal regression26. Ultrasound image attributes of bovine and equine CL reflected luteal and plasma progesterone content, and histomorphological characteristics of the CL18,27,28.
The role of the CL in regulating the development of ovarian follicular waves in women is not known. The objective of the present study was to characterize changes in the form and function of the CL during an IOI using serial high-resolution transvaginal ultrasonography, ultrasonographic image analysis and endocrine profiling. We hypothesized that changes in luteal area, progesterone concentrations, estradiol concentrations, and ultrasonographic image attributes would be detected during development and regression of the CL in women. We further hypothesized that changes in luteal area, endocrine production and ultrasonographic image attributes would differ between women with two versus three waves of follicular development.
Fifty women participated in a study to characterize ovarian follicular wave dynamics during the menstrual cycle11,12. Luteal image data and serum progesterone and estradiol-17β levels collected from these 50 women were used in the present study to evaluate the growth and regression of the CL during the IOI. Participants were assessed, by history and physical examination, to be healthy women of a mean reproductive age of 28 years (range, 19–43; SD ± 6.9). Women who were not eligible to participate included those who were pregnant at the time, had been pregnant or lactating 6 months prior to initiating study procedures, had used hormonal contraception 3 months before enrolling, had a history of irregular menstrual cycles, were taking medication(s) known or suspected to interfere with reproductive function, or were planning surgery during the study. The study protocol was approved by the Institutional Review Board of the University of Saskatchewan. Informed consent was obtained from all participants before initiating study procedures.
Each participant underwent daily transvaginal ultrasonographic evaluation of her ovaries for one IOI. The luteal phase was defined as the interval from the day of the first ovulation to the first day of menses. The follicular phase was defined as the interval from the first day of menses to the day of the subsequent ovulation. Scans were initiated 12 days after menses (i.e. before the first ovulation) and were continued until 3 days after the second ovulation. Ovulation was defined as the disappearance of a large follicle (> 15 mm) that had been identified by ultrasonography on the previous day, and was confirmed by the subsequent visualization of a CL9,10,29. Luteal glands were characterized as intraovarian structures with a thickened, irregular wall, with or without a hypo-echogenic central fluid-filled area at the site of the former pre-ovulatory follicle29 (Figure 1). High-resolution Ultra-mark 9 and ATL HDI 5000 ultrasound machines (Philips Medical Systems, Bothell, WA, USA) with 5–9-MHz multi-frequency convex array transducers were used to image the CL. Ultrasonographic examinations were performed by a single investigator (A.R.B.) approximately 90% of the time. A second investigator (R.A.P.) was available when the primary sonographer was not present.
Ultrasonographic images of the CL were obtained each day following ovulation until the CL was no longer visualized. Luteal image data were acquired via digital image transfer into a custom digital database. In addition, all ultrasonographic examinations were recorded onto S-VHS videotape for retrospective review if necessary. A customized computer program (Synergyne©, Women’s Health Imaging Research Laboratory, SK, Canada) was used to tabulate the cross-sectional area of the CL (based on the shape of an ellipse) and mean NPV of the ultrasonographic image of the CL. Cross-sectional area of the CL was determined by outlining the external border of the CL and the internal border of the central fluid-filled cavity (CFFC), if present. The area of the CFFC, if present, was subtracted from the overall luteal area to calculate the area that represented luteal tissue. Luteal tissue area, CFFC area and total luteal area were plotted separately for illustrative purposes.
NPV is a quantitative measurement of the gray-scale level of the pixel elements that comprise the ultrasound image, and is quantified using values ranging from 0 (black) to 256 (white)30. The CL image, at its maximal diameter, was divided into four equal quadrants and a circular region was used to sample NPV from each quadrant. Sampled regions encompassed only luteal tissue, avoiding the ovarian stroma and fluid-filled areas in the CL. The mean NPV was calculated by averaging the NPV measurements from the four quadrants28.
Blood samples were drawn to determine serum progesterone and estradiol-17β concentrations every third day in a stratified manner among women so that each day of the IOI was represented. Blood was collected into a 7 mL clot-activated tube and allowed to sit at room temperature for 15–30 min. Blood was then centrifuged for 10 min at 700 g and the serum was drawn off and stored at −20°C. Sequential competitive fluorescence immunoassays (Immulite®, Diagnostic Products Corporation, Los Angeles, CA, USA) were performed to measure serum progesterone and estradiol-17β levels. The interassay coefficients of variation (r) were as follows: (i) progesterone: low = 10.8%, medium = 7.0%, high = 10.8%, and (ii) estradiol: low = 9.8%, medium = 5.6%, and high = 4.3%. The minimal detectable limits for progesterone and estradiol were 0.2 ng/mL and 15 pg/mL, respectively.
Profiles of luteal area, luteal NPV, serum progesterone and serum estradiol concentrations during the IOI were centralized to the day of the first ovulation and plotted for each woman. Women were characterized as having either two (n = 34) or three (n = 16) waves of follicular development during the IOI, as previously reported11,12. Luteal area, luteal NPV and endocrine data were normalized to the mean IOI for women with two follicle waves (27.4 ± 0.4 days) or three follicle waves (29.4 ± 0.6 days) and the mean profiles were plotted. Mean luteal lifespan was defined as the mean day at which the CL could no longer be detected ultrasonographically. The day of onset of luteal regression was defined as the first day that met both of the following criteria: (i) luteal area less than for each of the previous 3 days and (ii) luteal area continued to decrease thereafter.
Repeated measures ANOVA were performed to assess changes in mean luteal area, NPV, progesterone and estradiol concentration during the IOI (PROC MIXED, SAS Institute, Cary, NC, USA, 2002). Independent sample t-tests and analyses of variance using Scheffe’s Post Hoc Tests and SPSS Version 11 (SPSS, Chicago, IL, USA) were used to compare maximum mean luteal area, NPV, progesterone and estradiol concentrations between women with two versus three follicle waves.
The CL was detected ultrasonographically on the day of ovulation and during the luteal phase in 100% of women evaluated (50/50), and during the subsequent follicular phase in 90% of women (45/50). Corpora lutea were of two morphological types: those with a CFFC and those without. The overall mean CFFC area measurement was 86.5 mm2. CFFCs ranged in size from 2.8–605.4 mm2. Serial images of corpora lutea containing no CFFC, a small CFFC and a large CFFC taken at different times during luteal growth and regression are shown in Figures 2 to to4,4, respectively. A CFFC was visualized in the CL of 44/50 women evaluated (88%), while a CFFC was not visualized in the remaining 6/50 women (12%). Forty-four out of fifty women (88%) exhibited a CL containing a CFFC 2 days after ovulation, followed by 17/50 women (34%) 13 days after ovulation and one woman (2%) 27 days after ovulation (Figure 5a).
In 27/50 women (54%) the second ovulation occurred ipsilateral to the first ovulation and in 21/50 women (42%) the second ovulation occurred contralateral to the first (P > 0.05). In the remaining 2/50 women (4%), two follicles ovulated simultaneously on both ovaries.
Changes in the total mean area of the CL during the IOI were plotted; total CL area was then partitioned to represent luteal tissue area and CFFC area separately (Figure 5b). Mean total CL area and mean luteal tissue area increased for 4 days following ovulation reaching peak levels of 411.6 mm2 and 328.0 mm2, respectively. Mean total CL area and luteal tissue area then decreased for the remainder of the IOI (day effect: P < 0.0001) (Figure 5b). Mean CFFC area remained relatively constant from days 1 to 11 following ovulation, reaching a peak value of 121.6 mm2 on day 7 followed by a subsequent decline (day effect: P < 0.0001) (Figure 5b).
Luteal tissue area was greater in women with a CFFC compared to those without a cavity during the IOI (day effect: P < 0.0001; day*cavity effect: P = 0.04) (Figure 5c). Luteal tissue area increased for 4 days after ovulation to peak values of 332.0 mm2 and 230.3 mm2 in women with and without CL containing fluid-filled cavities, followed by a subsequent decline (Figure 5c).
Changes in total mean CL area were superimposed with the changes in the diameters of the largest follicles of each wave in women with two and three waves of follicle growth during the IOI (Figure 6). Regardless of the number of follicle waves observed, the CL increased in diameter for the first week after ovulation (days 5 to 6) and then began to regress before the emergence of the second follicle wave. The length of the luteal phase was not different between women with two (13.4 ± 0.2 days) and three waves of follicle growth (13.1 ± 0.4 days, P = 0.60). Likewise, no differences were detected in the mean luteal lifespan of women with two follicle waves (19.6 ± 0.8 days) versus three follicle waves (19.9 ± 0.7 days, P = 0.4).
Mean total luteal area, serum progesterone concentrations and serum estradiol concentrations in women with two versus three follicle waves are compared in Figure 7. There were no differences in the changes in luteal area and progesterone concentration over time between women with two and three waves (day effect: P < 0.0001; day*wave effect: P > 0.05, respectively). Serum progesterone concentrations reached peak levels of 15.0 ng/mL and 14.0 ng/mL in women with two and three waves 6 days after ovulation followed by a decline to 2.2 ng/mL and 2.6 ng/mL on the first day of menses (day 13).
There were no differences in serum estradiol concentrations between women with two versus three waves during the luteal phase (day effect: P < 0.0001; day*wave effect: P = 0.6). However, estradiol concentrations rose earlier in the follicular phase in women with two versus three follicle waves (day effect: P < 0.0001; day*wave effect: P = 0.007), as previously reported12. In women with two follicle waves, serum estradiol concentrations increased to 117.6 pg/mL 6 days after ovulation and decreased to 30.2 pg/mL on day 13. In women with three waves, estradiol levels peaked at 109.7 pg/mL on day 8 and then decreased to 26.1 pg/mL on day 13.
Mean luteal area was positively correlated with serum progesterone concentrations in women with two (r = + 0.88 and three (r = + 0.90) waves of follicular development during the IOI. Similarly, luteal area was positively correlated with estradiol concentrations in women with two (r = + 0.62) and three (r = + 0.67) follicle waves during the luteal phase. No differences in maximum luteal area, peak progesterone concentrations or peak estradiol concentrations were detected between women with two versus three waves (Table 1, P > 0.05). Likewise, the first day of maximum luteal area, onset of luteal regression, peak progesterone concentration and peak estradiol concentration were not different between groups (Table 1, P > 0.05).
Luteal NPV during the IOI decreased from days 0 to 11 followed by an increase from days 11 to 16 (day effect: P < 0.0001). However, no differences in luteal NPV were detected between women with two versus three follicle waves (day effect: P < 0.0001, day*wave effect: P > 0.05) (Figure 8).
Changes in luteal form and function were characterized using serial high-resolution transvaginal ultrasonography, gray-scale image analysis and serum endocrine profiling. Our hypothesis that changes in luteal morphology and endocrine secretion would be detected during the IOI was supported. Two morphological types of CL were observed following ovulation: those with, and without, a CFFC. Most CL contained a CFFC. The incidence of corpora lutea containing a CFFC was greatest immediately following ovulation and then subsequently declined. CFFCs were attributed to the leakage of blood into the follicular lumen following follicle rupture. The ultrasonographic detection of a CFFC in the CL was therefore interpreted to be a normal physiological event during the menstrual cycle.
We were able to detect the CL in all women on the day of ovulation and throughout the luteal phase. The CL was also detected in 90% of women during the subsequent follicular phase. The side on which the first corpus luteum developed did not appear to influence the side on which the next ovulation occurred, as documented in previous studies31–33. Luteinization occurred for the first 6 days following ovulation in association with increasing levels of serum progesterone and estradiol. Thereafter, the CL regressed in association with decreasing concentrations of progesterone and estradiol. Although the regressing CL was present in the follicular phase, it did not appear to be functional as indicated by basal levels of serum progesterone and estradiol. Luteal area was highly correlated with progesterone concentrations during the IOI. Luteal area and estradiol concentrations, however, were not as well correlated. The source and role of luteal phase estradiol production is not fully understood. Given the new knowledge of follicular waves during the menstrual cycle, it is plausible that luteal phase estradiol levels may have a follicular, rather than luteal, origin.
Quantitative changes in luteal echotexture were reflective of changes in the morphological and physiological status of the CL in women, as previously documented in domestic animal species18,26–28. A decrease in luteal NPV occurred during luteal development in association with an increase in luteal area, progesterone and estradiol concentrations. The subsequent increase in NPV during luteal regression occurred in association with a decrease in luteal area, progesterone and estradiol concentrations. Decreased NPV during luteinization was attributed to increased vascularization of luteal tissue and a corresponding decreased tissue density. Increased NPV during luteolysis was attributed to decreased vascularization and replacement of luteal tissue with fibrous connective tissue, reflective of increased tissue density.
Our hypothesis that differences in luteal area, NPV and serum progesterone concentrations would be detected between women with two versus three follicle waves was not supported. The profiles of luteal area, NPV and progesterone concentration during the IOI did not differ in women with two versus three waves. Similarly, luteal phase estradiol concentrations did not differ in women with two versus three follicle waves, as previously documented12. However, an earlier rise in follicular phase estradiol concentrations in women with two waves occurred in association with an earlier emergence of the dominant ovulatory follicle12. No differences in luteal phase length, luteal lifespan, maximal luteal area, peak progesterone or peak estradiol concentrations were observed between the two groups of women. Similarly, the first day of maximal luteal area, progesterone concentration, estradiol concentration and the day of onset of luteal regression were not different in women with two and three follicle waves.
Our comparison of luteal endpoints between women with two versus three waves of follicular development does not correspond directly to those previously documented in domestic animals23. The regression of anovulatory dominant follicles during the bovine estrous cycle has been shown to occur through negative feedback effects of luteal progesterone on LH pulse frequency21,22. Periodic development of anovulatory follicle waves during the bovine estrous cycle continued until the CL regressed23. In both two and three wave animals, the CL regressed after emergence of the ovulatory wave, and before the emergence of the subsequent wave. The emergence of a third wave was associated with a longer luteal phase, and the viable dominant follicle present at the time of luteolysis became the ovulatory follicle23. Luteal regression occurred later in animals with three versus two follicle waves23,24 and progesterone levels remained elevated longer in three versus two wave animals25. Progesterone levels continued to rise during the bovine estrous cycle until the late follicular phase, at which time they dropped dramatically and the dominant follicle present ovulated22. In contrast to the bovine species, the CL in women regressed and progesterone concentrations decreased much earlier, in the mid-luteal phase of the menstrual cycle. That is, the second (ovulatory) wave emerged after luteal regression in women with two follicle waves. Similarly, the second (anovulatory) wave and third (ovulatory) wave emerged after luteal regression in women with three follicle waves. Therefore, waves of anovulatory follicular development in women did not continue until luteal regression, and the follicle present at the time of luteolysis did not become the dominant ovulatory follicle.
The differences observed in the growth and regression of the CL during the human menstrual cycle compared to that documented in the bovine estrous cycle could be due to two factors. Firstly, it is possible that the CL may act in a species-specific manner to influence the development of ovarian follicular waves. Major and minor waves of follicular development occur during the human menstrual cycle12 and equine estrous cycle15, but not during the bovine estrous cycle13. Therefore, the equine species may be a more appropriate model for studying the hormonal mechanisms underlying follicular and luteal dynamics in women34. The role of luteal progesterone and estradiol in regulating major and minor follicle wave dynamics in mares, however, has not yet been elucidated. Secondly, it is plausible that we have not yet been able to detect the precise mechanisms by which the CL regulates follicular growth and regression in women. The inability to obtain frequent (i.e. at least twice daily) blood samples from women in the study may have led to difficulty in detecting precise changes in luteal progesterone and endocrine secretion during the IOI.
In summary, serial high-resolution ultrasonography, ultrasonographic image analyses and serum endocrine profiling were shown to be effective methods for evaluating the morphology and function of the human CL during the menstrual cycle. We anticipate that the results of this study will provide ultrasonographic descriptions of luteal morphology, which will aid improvements in the clinical detection rate of corpora lutea in women during natural menstrual cycles and controlled ovarian hyperstimulation cycles, and provide a solid foundation for the detailed study of ovarian dynamics in women. Continued research should focus on comparative aspects of luteal function in women and domestic animal species, in order to determine whether species-specific luteal mechanisms exist to regulate the ovarian follicular wave phenomenon. Increased knowledge about luteal function during the spontaneous menstrual cycle would provide insight into the pathophysiology underlying luteal phase defects and/or recurrent miscarriage.
The authors would like to thank the research volunteers, whose participation was invaluable for the completion of this study. Appreciation is also expressed to Dr Norman Rawlings and Susan Cook for their expertise in endocrine immunoassays. Funding for this work was provided by the Canadian Institutes of Health Research.