Chronological aspects of ovulation cues: ultrasound, hormonal and other indirect cues

Chronological aspects of ultrasonic, hormonal, and other indirect indices of ovulation "

René Ecochard^a,*Hans Boehringer^bMuriel Rabilloud^aHenri Marret^c

Translation from English by Delphine Olive

Objective Improving ovulation prediction in normal cycles.

Design Collection of women's characteristics and menstrual cycles. Monitoring and analysis of temporal relationships between several ovulation indicators: transvaginal ultrasound, cervical mucus, temperature, etc.  basal, luteinizing hormone in urine and ratio of estrogen to urinary progesterone metabolites.

Installation Each of the eight natural family planning clinics was asked to study 12 women for at least three cycles.

Population One hundred and seven fertile women with normal cycles, aged 18 to 45.

Methods Daily measurements of luteinizing hormone, follicle-stimulating hormone, oestrone-3-glucuronide and pregnanediol-3α-glucuronide in urine. Temperature recording  basal and cervical mucus analysis. Transvaginal ultrasound of the ovaries.

Main results measured Discrepancies between the expected day of ovulation based on luteinizing hormone peak or ultrasound and the expected days based on other ovulation indices.

Results Ultrasound proved ovulation in 283 of 326 cycles. The mean interval between luteinizing hormone peak and ultrasound proof was less than a day (+0.46), but expected ovulation dates based on luteinizing hormone levels were observed early in almost 10 % of cycles or late in 23 % of cycles. A rise in temperature was observed  basal in 98 % of cycles. The symptom of peak cervical mucus, the rapid fall in urine levels and the initial rise in luteinizing hormone were all close to ultrasound evidence of ovulation in over 72 % of cycles.

Conclusions For reasons of precision and practicality, the symptom of peak cervical mucus, the ratio of metabolites in urine and the initial rise in luteinizing hormone are probably better clues to ovulation than the luteinizing hormone peak.

a Biostatics Unit, Department of Medical Information, Hospices Civils de Lyon, France
b Quidel Corporation, San Diego, USA
c Department of Obstetrics and Gynaecology, Hospices Civils de Lyon, France
* Correspondence: Dr R. Ecochard, Département d'Information Médicale, 162 avenue Lacassagne, 69424 Lyon Cedex 03 - France.

INTRODUCTION

There is a plethora of techniques claiming to predict ovulation^1,2 most of which are indirect methods. Laparoscopy is the reference direct method for observing ovulation, but it is technically difficult to perform and unrealistic to use on a daily basis. This is why new techniques for predicting ovulation are often evaluated in relation to the mid-cycle luteinizing hormone peak or ultrasound research.

The latter two references remain imperfect. What's more, they are not equivalent. Peak luteinizing hormone has been extensively evaluated, compared with direct laparoscopy.³ but it remains an indirect indicator of the expected day of ovulation (expected date of ovulation based on luteinizing hormone) and is often subject to biological variations.⁴. On the other hand, and although it is less precise than laparoscopy^5-9Ultrasound detection of the day of ovulation (ultrasound) is a direct method. However, it has the disadvantage of presenting the risk of subjective interpretations of ovarian morphology by ultrasound.

The choice of one of the two methods is crucial, particularly for assessing the length of the fertile period.¹⁰ or test new family planning methods^11-13. It is therefore necessary to systematically compare the expected date of ovulation according to luteinizing hormone levels and ultrasound as current reference days, as well as to study the temporal shifts between each method and other indirect indices of ovulation.

In the mid-1990s, a bank of first micturition urine specimens, collected on rising, was established by Quidel Corporation, San Diego, California, USA. The aim of the present study was to evaluate the discrepancies between the expected ovulation date based on luteinizing hormone and ultrasound, and to describe the temporal correlations between these two elements and several other indirect ovulation indices.

METHODS

The study protocol was written for a multi-center collaborative study under the auspices of Claude Bernard University, Lyon, France. It was approved by the Comité consultatif de protection des personnes dans la recherche biomédicale de Lyon. Subjects were recruited from eight natural family planning clinics located in Aix-en-Provence, Dijon, Lyon (France); Milan and Verona (Italy); Dusseldorf (Germany); Liège (Belgium); and Madrid (Spain). Each clinic was asked to study 12 or more women for at least three menstrual cycles. Women were informed of the purpose of the research and were free to withdraw from the study at any time. All participants gave written informed consent, and the research procedures were conducted in accordance with the code of ethics for human experimentation established by the Declaration of Helsinki.

To be included in the study, the women selected had to be settled and apparently healthy, aged 19 to 45 inclusive, with menstrual cycles of 24 to 34 days in length inclusive, and with experience of natural family planning methods (temperature  and signs of cervical mucus). The following potential participants were excluded: women with frequent anovulatory cycles; women undergoing hormonal response stimulation programs for fertility reasons; women undergoing oral, transdermal or other hormonal contraceptive programs; women undergoing a medical abortion program (< 3 months previous); and women undergoing hormone replacement therapy. Women with abnormal cycles (polycystic ovary syndrome or lutein deficiency), women with a history of fertility problems, who had undergone hysterectomy, tubal ligation or pelvic inflammatory disease, as well as runners, breastfeeding or postpartum women (< 3 months prior) were also excluded. In the end, one hundred and seven women were recruited.

Research

Temperature  basal body temperature was taken daily on waking, before any activity, and recorded on a special individual graph containing the date, day of cycle and any condition that might affect temperature (e.g. stress, illness or insomnia). The nadir of basal body temperature has been suggested as an indicator of ovulation (expected date of ovulation based on basal body temperature).  basal) and defined as the lowest point of the temperature curve, observed at the base of the rise corresponding to the hyperthermic phase.^14-15. We have also adopted the temperature-averaging method to determine the change between the hypothermic and hyperthermic phases.^16,17 : the average temperature over six consecutive days was calculated and the change (rise in temperature, fall in temperature, etc.) was compared.  identified as soon as a temperature was 0.2-0.3°C above this average³.

Cervical mucus at the vaginal entrance was assessed two or three times a day to note the sensation (dry, moist, wet, slippery) and consistency (sticky, creamy, elastic) on the temperature graph.  basal. The last day of clear, elastic and/or lubricating mucus has been called the peak mucus symptom.^18,19 hereafter referred to as the expected day of ovulation according to cervical mucus.

Each morning, urine samples were collected and prepared in three to five 10-12 mL aliquots in tubes containing gentamicin sulfate. Tubes were frozen at -20°C until analysis. On a daily basis, subjects were asked to perform two rapid tests with their morning urine: the Bluetest Ovulation test (Quidel) and the Conceive luteinizing hormone test (Quidel). The results were as follows  of these two tests should be noted on the temperature graph.  basal. This made it possible to detect the onset of luteinizing hormone rise and determine the appropriate day to start transvaginal ultrasound. Frozen urine specimens were tested, using a fluorimetric enzyme-linked immunosorbent assay (Delfia), for the hormones oestrone-3-glucuronide (E1-3-G), pregnanediol-3α-glucuronide (Pd-3α-G), follicle-stimulating hormone and luteinizing hormone. All samples from the same subject were tested in duplicate in the same assay and results adjusted for creatinine. Inter-assay variations were negligible; all urine samples were thawed once and analyzed in a single session. Intra-test variations were calculated as relative differences: (|value1-value2|)/mean value. These variations were 5.7 %, 6.8 %, 7.9 % and 8 % for Pd-3α-G, E1-3-G, luteinizing hormone and follicle-stimulating hormone respectively.

The initial rise in luteinizing hormone in urine of the first micturition (luteinizing hormone-IR) was determined as the first value three times higher than the average of the previous six baseline values. The luteinizing hormone peak in urine, which is close to ovulation (expected date of ovulation based on luteinizing hormone), was defined as the highest value, preceded and followed by lower values, and at least three times higher than the mean of the baseline values. When two peak values of identical magnitude occurred, the first was selected. Furthermore, the urinary level of E1-3-G/Pd-3α-G having been proposed as a means of determining the day of ovulation^{1, 20}We have calculated this rate using Baird's algorithm, which is an improvement on Royston's algorithm.²⁰.

Transvaginal ultrasound was used to monitor each ovulation. This was to be done as soon as fertile mucus appeared and/or luteinizing hormone rose, as detected by the rapid tests. Ultrasound was first performed daily until 16 mm follicles were observed, then daily until ovulation was proven. The day of ovulation (according to ultrasound, or JO-echo) was defined as the 24-hour interval between the visualization of a mature follicle on an image and the appearance of a ruptured follicle or an early corpus luteum, and/or free fluid in the cul-de-sac, on the following images⁵. More specifically, the signs of luteinization and/or follicle rupture visible by ultrasound were either a disappearance or a change in ultrasound size, shape or density.².

Statistical analysis

Mean hormone concentration values during the menstrual cycle were expressed as a percentage of maximum values and studied graphically with the JO-echo as the central point.²¹. Temporal correlations between expected ovulation date based on luteinizing hormone and JO-echo were studied both graphically and by Wilcoxon signed ranks test for paired samples. A systematic review of cycle data was carried out as soon as discrepancies were identified. Other indicators (basal temperature, Pd-3α-G, expected ovulation date according to the ratio) were then taken into account to confirm the ultrasound indications. First, we compared temporal correlations between expected ovulation date according to luteinizing hormone or ultrasound and other indirect indicators on the graph ; then, in a second step, we compared the proportion of cycles within two days of the expected ovulation date according to luteinizing hormone and ultrasound; then, in a third step, we compared the standard deviations of the intervals between the indirect indicators and the expected ovulation date. according to luteinizing hormone or ultrasound, respectively. All analyses were performed using S-PLUS statistical software (MathSoft, Inc.). Statistical significance was defined as P < 0,05 in all analyses.

RESULTS

Table 1 shows the characteristics of the selected subjects. The mean age of the women was 33.1 years (standard deviation 5.9, range 19-45), with a median of two previous pregnancies (range 0-6). The total number of cycles was 326, and the mean cycle length was 28.46 days (SD 3.55, range 22-48). On average, each woman contributed around three menstrual cycles. To be more precise: 22, 74, 5 and 6 women contributed four, three, two and one cycle(s) respectively. During the study, four pregnancies occurred; these were diagnosed by hCG levels in urine at the end of the cycles, then confirmed by ultrasound. Temperature curves  were determined either as a two-phase curve or as a single-phase curve using the temperature averaging method.^16,17. Overall, 310 cycles (95 %) showed biphasic temperature curves; the remaining 16 were monophasic. Our further analysis was restricted to menstrual cycles for which a JO-echo (ultrasound confirmed ovulation date) could be identified. In fact, in 28 of the 326 menstrual cycles (9 %), the first ultrasound search had been performed too late (the follicle had already ruptured), and we did not obtain ultrasound confirmation of ovulation in a further 15 cycles, leaving 283 cycles for further analysis (87 %). Luteinizing hormone-IR was identified in 273 of the remaining 283 cycles (96.5 %), expected ovulation date based on luteinizing hormone and expected ovulation date based on ratio in 272 cycles (96.5%). Temperature rise  was identified in 277 cycles (97.9 %), although the nadir of temperature  was identified in all cycles. The expected ovulation date based on cervical mucus was detected in 215 cycles (76 %).

Figure 1 shows the mean changes in the concentrations of four ovarian and pituitary/pituitary hormones over the 283 menstrual cycles, taking JO-echo as a reference. The pre-ovulatory E1-3-G rise began five to six days before ovulation. Another lower post-ovulatory peak was observed 6 to 11 days after ovulation. The concentration of Pd-3α-G increased rapidly before ovulation and reached its plateau around seven days after ovulation. Both follicle-stimulating hormone and peak E1-3-G were observed on the day of ovulation, followed by peak luteinizing hormone. An earlier, lower rise in follicle-stimulating hormone began before menstruation; the concentration peaked at the beginning of the cycle, followed by a decrease up to two days before the next ovulation.

As shown in figure 1, the luteinizing hormone curve peaked on the day following the JO-echo. Figure 2 shows a histogram distribution of the time lags between predicted ovulation date according to luteinizing hormone and JO-echo. The distribution of the predicted ovulation date according to luteinizing hormone is centered around the JO-echo. However, the mean delay between predicted ovulation date according to luteinizing hormone and JO-echo was +0.46 days, which is significantly different from 0 (P < 0,01). In 184 cycles out of 272 (67.6 %), the luteinizing hormone peak (predicted ovulation date based on luteinizing hormone) occurred from one day before to one day after the JO-echo.

Table 1. Characteristics of selected participants. Before entering the study, the 107 women completed an intake form that collected key socio-demographic data, as well as obstetric, gynecological and contraceptive history.

Variables	n (%)	Average	[Standard deviation]	Minimum	Median	Maximum
Age (years) Age at menarche (years) Size (m) Weight (kg) BMI (kg/m²) Previous pregnancies 0 1 > Loss of previous pregnancies Regular smoker Employed outside the home Sports activity Stress during the study	36 (33,6) 15 (14) 56 (52,4) 12 (11,2) 12 (11,2) 57 (53,3) 31 (29) 11 (10,3)	33,13 13,18 1,64 57,31 21,24 1,81	[5,86] [1,64] [0,07] [7,90] [2,55] [1,69]	19 9 1,47 37 17,12 0	33,23 13 1,64 56 20,70 2	45 17 1,80 80

 

Figure 1. Mean urinary concentrations of E1-3-G, Pd-3α-G, follicle-stimulating hormone (FSH) and luteinizing hormone (LH) in relation to the day of ovulation as determined by ultrasound (US-DO or JO-echo). Hormone concentrations were first expressed in the most common units (ng/ml, μg/ml, mIU/ml and mIU/ml, respectively), adjusted for creatinine, then transformed into % of the maximum mean value for each hormone.

In a further 26 cycles (9.6 %), the luteinizing hormone peak preceded the JO-echo by more than a day. These 26 cycles were often long cycles with a minor first luteinizing hormone peak and delayed ovulation, concomitant with the main peak (see the typical profile shown in Fig. 3a).

  

Figure 2. Distribution of temporal shifts between the day of ovulation as determined by peak luteinizing hormone (predicted ovulation date based on luteinizing hormone) and as determined by ultrasound (JO-echo or US-DO). The vertical axis shows the actual number of cycles. n = the number of cycles for which both indices were available.

  

Figure 3. Typical cases of discrepancies between the ovulation date predicted by the luteinizing hormone and the JO-echo (US-DO). (a) Low early and late luteinizing hormone peaks; (b) Single late peak of luteinizing hormone.

In 62 cycles (22.8 %), the predicted ovulation date based on luteinizing hormone occurred more than a day after JO-echo. In these cycles, luteinizing hormone began to rise before JO-echo but continued to rise after ovulation (see cycle shown in Figure 3b). In most cases, a systematic review of cycle data confirmed JO-echo using the expected ovulation date based on ratio, the expected ovulation date based on basal body temperature and the expected ovulation date based on cervical mucus. Very often, prolonged luteinizing hormone peaks occurred in parallel with a slower increase in pregnanediol.

We also studied the distribution of intervals between each indirect ovulation index (expected ovulation date according to cervical mucus, expected ovulation date according to basal body temperature, expected ovulation date according to ratio, temperature rise, etc.).  basal and initial rise in luteinizing hormone) and each reference day (predicted ovulation date based on luteinizing hormone and JO-echo). The results are shown in Table 2 and Figure 4. Immediately after the initial luteinizing hormone rise, which is naturally close to the luteinizing hormone predicted ovulation date, the index closest to the luteinizing hormone predicted ovulation date was the ratio predicted ovulation date (estrogen versus progesterone metabolites in urine): the difference was one day or less in 76.2 % cycles. The expected date of ovulation based on cervical mucus, the expected date of ovulation based on the ratio and the initial rise in luteinizing hormone were all close to the JO-echo in over 72 % of cycles. Standard time lags between the different indices and the JO-echo were 1.6, 2.2 and 2.4 days for cervical mucus ovulation date, luteinizing hormone initial rise, and ratio predicted ovulation date, respectively.

Table 2. Quality of ovulation indexes

Caption:
Legend :
CM-EDO Expected day of ovulation based on cervical mucus.
BBT-EDO Expected day of ovulation based on basal body temperature
LH-EDO Expected day of ovulation according to luteinizing hormone
R-EDO expected day of ovulation based on ratio
BBT rise rise in basal body temperature
LH-IR initial rise in luteinizing hormone

  

Figure 4. Distributions of temporal shifts between the day of ovulation as determined by reference methods (expected ovulation date based on luteinizing hormone and JO-echo) and that determined by five indirect methods (expected ovulation date based on cervical mucus, expected ovulation date based on temperature, etc.).  basal body temperature, expected ovulation date according to ratio, basal body temperature rise and initial luteinizing hormone rise). The vertical axis shows the actual number of cycles. n = the number of cycles for which both indices were available.

DISCUSSION

Composite profiles of daily urinary concentrations of E1-3-G, Pd-3α-G, follicle-stimulating hormone and luteinizing hormone have already been described by various authors, in hypofertile or normally fertile women.^20-22. However, descriptions of their temporal correlations with JO-echo have often been based on a limited number of cycles. To date, our study presents the largest number of cycles for which ultrasound indices (JO-echo), clinical indices (ovulation date based on cervical mucus, ovulation date based on temperature  basal) and hormonal indices (expected ovulation date based on luteinizing hormone and ovulation date based on ratio) were determined. Based on a relatively large number of normally fertile women, this study is a valuable source of data.

The ultrasound, clinical and hormonal criteria for ovulation were planned and proved to be extremely homogeneous, which also proves the value of this study. In fact, all the ultrasound research centers strictly adhered to the JO-echo criteria for detecting and determining ovulation. In addition, urine samples were centrally analyzed for hormone concentrations, reducing inter-dose variability.

However, despite all precautions, no JO-echo could be determined for 13 % of the cycles, due to late search (9 %) or lack of attendance (4 %). Scheduling several ultrasound searches per day at appropriate times could have reduced this percentage, but such a protocol seems to be more easily achievable in institutions dedicated to research rather than in family planning centers. Indeed, in the context of our study, requesting more research would have prevented some centers from participating, as it would have disrupted their usual pattern of operation as intervention centers; moreover, it would have made the recruitment of normal subjects more difficult due to the additional constraints.

Comparison of the urinary hormone profiles we obtained with previously published plasma hormone profiles^21,23 led us to note a number of similarities: 1. the urinary peak of E1-3-G and the peak of plasma estradiol occurred on the day before the peak of luteinizing hormone hinge 2. pregnanediol levels and plasma progesterone increased in conjunction with the luteinizing hormone peak; 3. the urinary E1-3-G peak and plasma estradiol peak preceded the luteinizing hormone peak and were of longer duration; 4. the luteinizing hormone peak lasted three days or more in urine and plasma. The luteinizing hormone peak lasted three days or more in urine and plasma. Nevertheless, there are some discrepancies between urinary and plasma results: in our study, the mid-cycle urinary follicle-stimulating hormone peak occurred on the same day as the urinary E1-3-G peak, one day before the luteinizing hormone peak, which contrasts with the concomitant plasma peaks of follicle-stimulating hormone and luteinizing hormone. This may be due to the fact that the follicle-stimulating hormone peak was narrow and the luteinizing hormone peak was wide: luteinizing hormone continued to rise after ovulation in several cycles.

Ultrasound research has been proposed as a reference method for predicting the day of ovulation². However, due to the technical parameters involved and the high level of commitment required of the subjects, its use is mainly accepted by women with major fertility problems, or by those taking part in a medically assisted reproduction program, ideally artificial insemination. For the vast majority of practitioners and women interested in ovulation detection, less costly, less time-consuming and much easier-to-use methods are needed on a day-to-day basis; these should be compared with the most effective direct method (i.e. ultrasound).

Among the hormonal methods we used to predict the day of ovulation, the measurement of Pd-3α-G concentration seemed less reliable than the other methods due to its late peak and distance from the JO-echo, but it remains potentially interesting for identifying the post-ovulatory infertile phase. Conversely, E1-3-G and follicle-stimulating hormone peaked on the day of ovulation, making them very good indicators. Furthermore, in line with the results of the WTO's real-life trials³the expected ovulation date based on the ratio was also reliable in 75 % cycles and, being based on a ratio of hormone levels, is not subject to variations in urine concentration and therefore does not require adjustment to creatinine.²⁴.

However, Pd-3α-G, E1-3-G and follicle-stimulating hormone tests are still expensive, less readily available as home kits and mainly performed by qualified laboratory staff. Their use is therefore still limited to assisted reproduction programs.. There is a strong demand for inexpensive home tests measuring the ratio of Pd-3α-G/E1-3-G levels.

Our results showed that the JO-echo and the expected ovulation date based on luteinizing hormone can be considered equivalent on average. Nevertheless, if we consider the JO-echo +/- 1 day as a fair estimate of the day of ovulation, the initial rise in luteinizing hormone appeared to be more effective than the luteinizing hormone peak. per se (77 % versus 67 % cycles). This corroborates previous assertions on the importance of the initial rise in luteinizing hormone in estimating the day of ovulation.^25,26. Taken together, detection of the initial rise in luteinizing hormone and monitoring of the luteinizing hormone peak can offer a fairly reliable strategy for predicting ovulation in almost 90 % of normal cycles, including long cycles. This combined method can therefore be recommended for planning a natural or medically-assisted pregnancy, but not for avoiding pregnancy.²⁷. What's more, the luteinizing hormone test is relatively affordable and easily performed at home.

Nadir or rising temperature  basal timing did not appear to be satisfactorily reliable, as the former often preceded the JO-echo, while the rise often followed the JO-echo or luteinizing hormone-expected ovulation date by more than two days (JO-echo +/- 1 day in almost 13 % vs. luteinizing hormone-expected ovulation date +/- 1 day in almost 23 % cycles). An error rate of 6.8 to 9.5 % seemed inherent in using this method to determine the ovulatory status of healthy women.¹⁴. Conversely, the symptom of peak cervical mucus was much closer to the JO-echo or the expected ovulation date according to luteinizing hormone (JO-echo +/- 1 day in almost 75 % of cycles compared to the expected ovulation date according to luteinizing hormone +/- 1 day in almost 70 % of cycles). The advantage of self-identification of the cervical mucus peak sign over temperature self-monitoring  has already been demonstrated¹⁹.  So, as already stated^19,20,23The cervical mucus peak sign can be an interesting clinical sign to detect ovulation. The method can be conveniently used by women at home for moderately rigorous or simple family planning.

CONCLUSION

The results of this study, as well as those previously published, clearly establish a very high correlation between the expected ovulation date based on luteinizing hormone and the JO-echo. Nevertheless, in a non-negligible number of cycles, these two indicators did not correspond. We therefore believe that the initial rise in luteinizing hormone may be a better indicator of ovulation than the luteinizing hormone peak itself. Furthermore, our research confirms the interesting role of the cervical mucus peak sign as well as that of the ratio of estrogen to progesterone metabolites in urine as indicators of ovulation. However, these two methods differ considerably in terms of their purpose, cost and set-up parameters.

 

British Journal of Obstetrics and Gynaecology
August 2001, Vol. 108, pp. 822±829

original article: https://obgyn.onlinelibrary.wiley.com/doi/full/10.1111/j.1471-0528.2001.00194.x?fbclid=IwAR3IMq8jk9OhZEYOvnCo3vliY5lUouG2s1zv_J0zbXT9LAZn6VjcUcU1-d0 " 

Thanks

The authors would like to thank Drs S. Dubus, A. Leduy, I. Ecochard, M. Grisard Capelle, E. Barranco, M. Barbato, S. Girotto and M. Gimmler of the natural family planning clinics, as well as all the women who participated in this study. We also thank Dr J. Iwaz, PhD, scientific advisor, for his suggestions and criticism of the manuscript. This study was partially supported by Quidel Corporation, San Diego, California, USA.

References

1. Royston P. Identifying the fertile phase of the human menstrual cycle. Stat Med 1991;10:221±240.

2. Moghissi KS. Ovulation detection. Reprod Endocr 1992;21:39±55.

3. WHO. Temporal relationships between ovulation and defined changes in the concentration of plasma estradiol-17β luteinizing hormone, follicle-stimulating hormone, and progesterone. Am J Obstet Gynecol 1980; 138:383±390.

4. Garcia JE, Jones GS, Wright Jr GL. Prediction of the time of ovulation. Fertil Steril 1981;36:308±315.

5. Queenan JT, O'Brien GD, Bains LM, Simpson J, Collins WP, Campbell S. Ultrasound scanning of ovaries to detect ovulation in women. Fertil Steril 1980; 34:99±105.

6. Marinho AO, Sallam HN, Goessens LK, Collins WP, Rodeck CH, Campbell S. Real time pelvic ultrasonography during the periovulatory period of patients attending an artificial insemination clinic. FertilSteril 1982; 37:633±638.

7. Vermesh M, Kletzky OA, Davajan V, Israel R. Monitoring techniques to predict and detect ovulation. Fertil Steril 1987; 47:259±264.

8. Shoupe D, Mishell DR, Lacarra M, et al. Correlation of endometrial maturation with four methods of estimating day of ovulation. ObstetGynecol 1989; 73:88±92.

9. Luciano AA, Peluso J, Koch EI, Maier D, Kuslis S, Davison E. Temporal relationship and reliability of the clinical, hormonal, and ultrasonic indices of ovulation in infertile women. Obstet Gynecol 1990; 75:412±416.

10. Wilcox AJ, Weinberg CR, Baird DD. Timing of sexual intercourse in relation to ovulation. N Engl J Med 1995; 333:1517±1521.

11. Royston JP, Humphrey SJE, Flynn AM, Marshall J, Zarzosa-Perez A. An automatic electronic device (Rite time) to detect the onset of the infertile period by basal body temperature measurements. Br J Obstet Gynaecol 1984; 91:565±573.

12. Ismail M, Arshat H, Pulcrano J, Royston P, Spieler J. An evaluation of Bioself 110 fertility indicator. Contraception 1989; 39:53±71.

13. Bonnar J, Flynn A, Freundl G, Kirkman R, Royston P, Snowden R. Personal hormone monitoring for contraception. Br J Fam Plann 1999; 24:128±134.

14. Hilgers TW, Bailey AJ. Natural family planning. II. Basal body temperature and estimated time of ovulation. Obstet Gynecol 1980; 55:333±339.

15. De Mouzon J, Testart J, Lefevre B, Pouly JL, Frydman R. Time relationships between basal body temperature and ovulation or plasma progestins. Fertil Steril 1984; 41:254±259.

16. McCarthy JJ, Rockette HE. A comparison of methods to interpret the basal body temperature graph. Fertil Steril 1983; 39:640±646.

17. Vollman RF. The Menstrual Cycle. Philadelphia: WB Saunders, 1977.

18. Billings EL, Billings JJ, Brown JB. Symptoms and hormonal changes accompanying ovulation. Lancet 1972; 1:282±284.

19. Hilgers TW, Abraham GE, Cavanaugh D. Natural family planning. I. The peak symptom and estimated time of ovulation. Obstet Gynecol 1978; 52:575±582.

20. Baird DD, Weinberg CR, Wilcox AJ, McConnaughey DR, Musey PI. Using the ratio of urinary estrogen and progesterone metabolites to estimate day of ovulation. Stat Med 1991; 10:255±266.

21. Collins WP. Biochemical indices of potential fertility. Int J Gynecol Obstet 1989; 1:35±43.

22. Brown JB, Blackwell LF, Homes J, Smyth K. New assays for identifying the fertile period. Int J Gynecol Obstet 1989; 1:111±122.

23. Leader A, Wiseman D, Taylor PJ. The prediction of ovulation: a comparison of the basal body temperature graph, cervical mucus score, and real-time pelvic ultrasonography. Fertil Steril 1985; 43:385±388.

24. Baird DD, McConnaughey DR, Weinberg CR, et al. Application of a method for estimating day of ovulation using urinary estrogen and progesterone metabolites. Epidemiology 1995; 6:547±550.

25. Singh M, Saxena BB, Rathman P. Clinical validation of enzyme immunoassay of human luteinizing hormone (hluteinising hormone) in the detection of the preovulatory luteinizing hormone (luteinising hormone) surge in urine. Fertil Steril 1984; 41:210±217.

26. Kesner JS, Knecht EA, Krieg EF, Wilcox AJ, O'Connor JF. Detecting pre-ovulatory luteinizing hormone surgesin urine. Hum Reprod 1998; 13:15±21.

27. Flynn AM. Natural family planning and the new technologies. Int J Gynecol Obstet 1989; 1:123±127.

Accepted February 5, 2001