- Email: [email protected]
Temporal relationships between indices of the fertile period
Vol. 39, No.5, May 1983 Printed in U.8A.
FERTILITY AND STERILITY Copyright 0 1983 The American Fertility Society
Temporal relationships between indices of the fertile period
World Health Organization Task Force on Methods for the Determination of the Fertile Period, Special Programme of Research, Development and Research Training in Human Reproduction. World Health Organization, Geneva, Switzerland
The intra- and interwoman variation in nine physiologic or biochemical indices of the fertile period has been studied over 58 menstrual cycles from 13 experienced users of the symptothermal method of family planning by periodic abstinence. The time and duration of a possible fertile period have been determined by five methods (symptothermal, cervical mucus, basal body temperature plus calendar calculation, defined changes in the concentration of estrone-3 -glucuronide and the ratio of estrone-3-glucuronide to pregnanediol-3a-glucuronide as determined by immunochemical tests on daily samples of early morning urine). The values were compared with a period of probable fertility (day of urinary luteotropin [LHJ peak - 3 to day of LH peak + 2). The duration of the possible fertile period by each method (mean ± standard deviation) was 13.4 (2.9), 11.9 (2.9), 11.8 (3.3),9.3 (2.2), and 10.9 (2.3) days, respectively, while the percentage of the probable fertile periods covered entirely by each approach was 98%, 91%, 90%, 83%, and 84%, respectively. The results warrant the initiation of clinical trials to ascertain the practical value of the individual or combined tests for family planning and the management of infertility. Fertil Steril 39:647, 1983
The true fertile period (FP) in women only occurs when there is a viable ovum. Ovulation is the crucial event during the menstrual cycle, and it has been calculated that the maximum survival time for the ovum is probably less than 48 hours. 1 Received September 13, 1982; revised and accepted December 22, 1982. Reprint requests: Professor W. P. Collins, Department of Obstetrics and Gynaecology, King's College Hospital, Denmark Hill, London SE5 8RX, United Kingdom. The Task Force Investigators for this study were William Collins, Ph.D., D.Sc., Janet Simpson, B.Sc., and Patricia Collins, B.Sc., Department of Obstetrics and Gynaecology, King's College Hospital, London, United Kingdom; Anna Flynn, M.R.C.O.G., and Maureen McGarry, Birmingham Maternity Hospital, Queen Elizabeth Medical Centre, Birmingham, United Kingdom; Patrick Royston, B.A., M.Sc., M.I.S., Division of Computing and Statistics, MRC Clinical Research Centre, Harrow, United Kingdom; and Jeffrey Spieler, M.Sc., and Alain Pinol, D.E.S.T., Data and Text Processing Services Unit, Division ofInformation Systems Support, World Health Organization, Geneva, Switzerland. Vol. 39, No.5, May 1983
After intercourse, the spermatozoa may have a longer life span (up to 6 days), and their fertilizing potential is dependent upon the quality of mucus in the female genital tract, which enables the gametes to meet in one of the oviducts. Hence, the potentially FP of a couple is the time from the first act of intercourse which may lead to pregnancy to the demise of the ovum. Accordingly, attempts have been made to identify three phases during each menstrual cycle: (1) an initial phase when there is only a low risk that intercourse could lead to conception; (2) a periovulatory phase when the chance of fertilization increases to a maximum and then rapidly falls; and (3) a phase of absolute infertility when the ovum is no longer viable. 1 In practice, there may be more than one apparent FP each menstrual cycle. 2 The application of radioimmunoassay (RIA) and pelvic ultrasonography to the daily monitoring of healthy women has established that the development of an ovarian follicle is associated World Health Organization Indices of the fertile period
647
with an increase in the concentration of estradiol (E 2) in peripheral venous plasma. 3 Furthermore, the rapidly increasing levels of this hormone in the peripheral circulation stimulates an increase in the production and changes in the characteristics of cervical mucus (CM) and subsequently initiates a surge in the release ofluteotropin (LH) from the pituitary.4 Consequently, defined changes in the concentration of E2 in serial samples of peripheral plasma or daily changes in the characteristics of CM have been used to indicate ovulation and the probable start of the FP. Other studies with labeled E2 and RIAs have shown that from 50% to 80% of the circulating hormone is excreted in the urine over a period of 24 hours and that estrone-3-glucuronide (E 1-3-G) is one of the principal metabolites. The concentration of this conjugate may be measured in diluted samples of early morning urine (EMU) without prior hydrolysis or extraction and characteristic changes used to estimate the start of the probable FP. 5 Alternatively, the likely start of each FP has been calculated from a knowledge of the previous menstrual history, e.g., the shortest length ofthe previous six cycles minus 18 days. 6 The resumption of oogenesis, 7 follicular rupture, and luteinization of the theca and granulosa cellss, 9 occurs after a noticeable increase in the level of circulating LH. For example, the results from a World Health Organization (WHO) multicenter studys indicated that a defined rise in the level of plasma LH (a 50% increase in the mean of three preceding values; sampling every 8 hours) was a good index of impending ovulation, which occurred from 24 to 56 hours later (mean, 32 hours). Moreover, LH is rapidly excreted in urine, and the time interval from the rise of the metabolite in EMU to follicular rupture as determined by ultrasonography is 20 to 44 hours, with a mean of 30 hours. 1o Luteinization, in terms of increased progesterone (P) in the peripheral circulation, immediately precedes ovulations, 11 and is associated with changes in the amount and physicochemical properties of CM. Hence, an appropriate rise in the concentration of plasma P or a change in the characteristics of CM12 may be used as reference points to calculate the end of the FP. The clearance ofP from the vascular compartment (approximately 25% over 24 hours) is slower than for E2 and may explain why there are limitations to the use of changes in the concentration of pregnanediol-3a-glucuronide (Pd-3a-G) to locate the end 648
World Health Organization Indices of the fertik period
of the FP. 5 However, the observation that basal body temperature (BBT) rises around the time of follicular rupture has been used to calculate the start of the postovulatory infertile phase of the ovarian cycle. 12, 13 The present study was prompted by results on the practical value of the CM 14,15 and symptothermal methods 16 to delineate the FP and the recent development of immunochemical tests based on the analysis of steroid glucuronides in EMU to achieve the same objective. 5 The aim was to compare (both within and between women) the time relationship between five different indices of the start and four of the finish of each FP. In addition, the duration of the FP was calculated by five methods (symptothermal, CM, BBT plus calendar calculation, and immunochemical tests based either on changes in the concentration of E 1-3-G in EMU or in the derived ratio E 1-3-G to Pd-3a-G). The results are compared in terms of the accuracy with which the FP is located and the time of abstinence required if the method had been used to avoid a pregnancy.
MATERIALS AND METHODS
The intention was to study at least ten healthy women for six consecutive menstrual cycles. All potential participants were informed of the purpose of the investigation and reassured that they would be free to withdraw at any time. The subjects were recruited from the Natural Family Planning Clinic at Birmingham Maternity Hospital, and physiologic indices of the FP were recorded in the home on a special form provided by WHO. Daily samples of EMU were collected, stored at - 15° C, and subsequently sent to King's College Hospital Medical School, London, for the analysis ofE 1-3-G, LH, and Pd-3a-G. The results were also recorded on forms provided by WHO. All forms (from Birmingham and London) were sent independently to WHO in Geneva for data processing. SUBJECTS
Thirteen experienced users of the symptothermal method of family planning were studied. They were between 25 and 33 years of age, and ten had had at least one child. For inclusion in the study, each woman had to have experienced regular menstrual cycles (25 to 35 days for six previous cycles) and not be receiving any form of Fertility and Sterility
chronic drug therapy or working night shifts. In addition, no subject had a history of or evidence of liver or kidney disease or dysfunction or had used any form of hormonal contraception or intrauterine device during the previous three menstrual cycles. All participants had to be willing to complete symptothermal charts· and collect EMU for six complete menstrual cycles. METHODS
The BBT was taken each morning, before getting out of bed. A mercury thermometer with an expanded scale was used. The presence and type of CM at the vulva was assessed according to sensation and described as dry, damp, or wet. In addition, each subject wiped the vulva with toilet paper before and after micturition, and recorded the sensation and appearance of the mucus. Observations were made throughout the day and an assessment was made immediately before retiring to bed at night. The concentrations of E l -3-G, LH, and Pd-3aG were measured by RIA without prior hydrolysis or extraction. The methods have been described and evaluated in detail. 17 DEFINITIONS
The following definitions were used. Day 1 of the menstrual cycle was the first day of menstruation, i.e., the day starting after midnight on which there was a noticeable flow of blood (requiring sanitary protection) through the vagina. The day of LH peak was the day of the highest arithmetic value for urinary LH between days 5 and 25 of the menstrual cycle. The probable FP was defined as the day of the LH peak - 3 to the day of the LH peak + 2, inclusive. A signal on day LH peak - 3 was accepted as indicating the start of the FP and on day LH peak + 2 as the beginning of the infertile phase. The calculated start of the FP was the day of the current menstrual cycle resulting from the calculation (number of days in the shortest of six previous cycles -18). The length of each new cycle was subsequently used in the calculation. The first day of mucus was the day on which there was a sensation of dampness (i.e., no longer dry) or the first visual appearance of mucus. The first day of fertile-type mucus was the first day that a sensation of wetness or slippery feeling was recorded (with or without the appearance of any kind of mucus) or the appearance of clear stretchy mucus. The day of peak Vol. 39, No.5, May 1983
mucus was the last day of the fertile-type mucus, which could only be determined retrospectively, i.e., on the next day. An EMU was defined as the first urine passed upon awakening in the morning. The days on which a sustained rise occurred in the concentration of E l -3-G and E l -3-GIPd-3a-G were determined by an adaptation of the cumulative sum (CUSUM) test. 18 All values were converted to IOg10 before analysis. A baseline was calculated as the mean of values on days 1 to 6 for each cycle studied. A reference level was calculated as baseline + 1 standard deviation (SD). A single value for the mean population SD was found to be adequate for each of the variables and was 0.30 for E l -3-G and 0.35 for the ratio, both in log units. The CUSUM test was performed by subtracting the reference level from each daily value and adding the result to the current CUSUM. A significant rise was recorded when the CUSUM reached a decision level, which was taken as 2 SD above the mean baseline value. The day of the defined rise (used to indicate the start of the FP) was the day on which the data first exceeded the reference level in the run leading to a significant CUSUM. The peak days for E l -3-G and the ratio were determined by starting with the rise day + 1 and taking the day with the highest value that was followed by three consecutive lower amounts. The third day of higher phase (postovulatorypremenstrual) BBT was also identified by a modified CUSUM technique. 18 The baseline was taken as the mean temperature of 8 days, starting at day 4, and the reference level as baseline plus 0.1 0 C for each cycle. The CUSUM showed a significant rise when it exceeded 0.25; however, this could occur before three high temperatures had been recorded. For standardization of the procedure, three consecutive temperatures greater than or equal to the reference value were required before the third of these was selected as the last day of the FP (BBT + 3). Days were excluded on which the subject recorded more than 1 day of illness, which elevated the temperature. If the time of taking the BBT was recorded as later than usual, the reading was reduced by 0.10 C for each additional hour.19 LIMITS OF THE FERTILE PERIOD
All indicators of the FP were timed relative to the day of the LH peak; for example, day -1 was World Health Organization Indices of the fertile period
649
Table 1. The Number of Subjects and Cycles Studied
Total
No. of subjects
No. of menstrual cycles/subject
7 1 2 2 1
6 5 4 2 1
No. of conceptional cycles a
13
(
No. of cycles excluded from analysisb
42 4 7 4 1
1 1
1 1
2
2
Total no. of cycles analyzed
58
alncluded for data processing. bExcluded because of missing samples of EMU.
the day preceding the peak, and day + 1 was 1 day after the peak. There was five potential markers of the first fertile day: F1, the shortest of the last six menstrual cycle lengths - 18; F2, the first day of any eM; F3, the first day of fertiletype eM; F4, the day of defined rise in E1"3-G; and F5, the day of a defined rise in E l -3-G/Pd3a-G. Four indices of the end of the FP were compared: L1, the day of peak E1"3-G + 4; L2, the day of peak mucus + 4; L3, the third day of high BBT (BBT + 3); and L4, the day of peak ratio +
peak + 2, (i.e., the probable FP). One-way analysis of variance was used to determine whether there were significant differences between women for any of the variables. Similar analyses were applied to the five methods of calculating the duration of the FP (D1 to D5). In particular, the proportion of cycles was calculated in which each method failed to cover the period of the day of LH peak -3 to day ofLH peak +2.
RESULTS
5. LENGTH OF THE FERTILE PERIOD
The duration of the FP was estimated by five methods: D1, symptothermal (later of L2 and L3 - earlier of F1 and F2 + 1); D2, eM (L2 - F2 + 1); D3, calculation and BBT (L3 - F2 + 1); D4, chemical test 1-E l -3-G (L1 - F4 + 1); and D5, chemical test 2-ratio of E l -3-G to Pd-3a-G (L5 - F5 + 1). D1, D2, and D3 represent variants of techniques used in natural family planning (NFP) programs, and D4 and D5 are immunochemical tests that might be considered as alternatives to D1, D2, or D3. STATISTICAL ANALYSIS
The purpose of the analysis was to examine the time intervals between five indices to predict the start of the FP (F1 to F5) and the day ofLH peak, and between four indices to detect the end (L1 to L4) from the same reference point. A signal was said to have occurred on a given day if the start or end of the fertile period was indicated by one of the methods. The cumulative proportion of cycles with each signal was plotted on probability paper against the day of the signal relative to the day of the LH peak. In this way, the number of failures and successes of each index was determined by reference to the day of LH peak - 3 to day of LH 650
World Health Organization Indices of the fertile period
The number of subjects recruited and menstrual cycles analyzed are shown in Table 1. The results from two cycles were excluded from the analysis because samples of EMU were missing for at least 5 consecutive days during the time of follicular development. An additional 40 samples of EMU were missing « 3% of total), but this limitation would not have affected the temporal relationship of any hormonal index by more than 24 hours.
Peak mucus +4 BBTshift +3
--. Mean and
observed n = 58
Peak ratio +5 Peak E.-3-G +4
First fertil. mucus Defined rise in Defined riee In First mucus Short.st cycle -18
•
-10
Figure 1
-5 .0 +5 +10 Days relative to LH peak
The means and 90% observed ranges for the days of nine signals relative to the day of the LH peak.
Fertility and Sterility
Table 2. Intra- and Interwoman Variation-One-Way Analysis of Variance for Nine Indices of the Start or Finish of the FP Relative to the Day of LH Peaka Variable Length of menstrual cycle Day of LH peak Shortest cycle length -18 First day of mucus First day of fertiletype mucus Day of defined rise in E l -3-G Day of defined rise in ratio Day of peak E l -3-G +4 Day of peak mucus +4 BBT rise +3 Day of peak ratio +5
Mean
SD
Minimum
Difference between women
Maximum
(P)
Witbinwoman SD
No. of cycles with signal withinFP
27·.7
2.7
22
35
0.0001
2.0
15.0 -7.0
2.6 2.5
10 -14
23 -1
0.0001 0.3
2.0 2.5
2
-6.6 -4.4
2.5 2.6
-12
-11
-2 0
0.0002 0.03
2.0 2.3
1 14
-4.9
1.9
-10
-1
0.3
1.8
5
-6.5
2.8
-14
-1
0.007
2.4
1
3.4
1.9
-4
9
0.09
1.7
5
4.3
2.2
-5
10
0.055
2.0
3
3.8 3.2
2.0 2.3
-1 -4
11
0.0008 0.7
1.6 2.3
4 8
6
aOverall means and ranges for 58 menstrual cycles and number of inadequate signals.
INDICES FOR THE LIMITS OF THE FERTILE PERIOD
The means and 90% observed limits for the nine indices are shown in Figure 1 relative to the probable period of fertility. The 90% limits were calculated as the 4th and 55th ordered values in the sample of 58 cycles. The numeric values for the means, SDs, and ranges of the indices, together with the lengths of the menstrual cycles and the days of the LH peaks, are listed in Table 2. The significance of difference for each variable between women, together with the SD within women, is also shown in Table 2. The withinwoman SD represents the variability of values from cycle to cycle in each subject. The cumulative percentage of cycles giving each signal used in NFP on days relative to the day of the LH peak is shown in Figure 2; and from those used in the immunochemical tests, in Figure 3. The number of cycles where each signal falls within the defined FP (Le., late indication of the start or falsepositive detection end of the FP) is listed in Table
2. DURATION OF THE FERTILE PERIOD
The apparent duration of the FP for all cycles as determined by the five combined indices (D1 to D5) is shown as frequency distribution curves in Figure 4. The means, SDs, and ranges of values as determined by each method, together with the significance of difference between intra- and inVol. 39, No.5, May 1983
terwoman variation, are shown in Table 3. The number of cycles where the calculated FP by each method did not entirely cover the defined FP ranged between 2% and 17%. DISCUSSION
This report describes a prospective study of the time intervals between indices used in NFP to locate the probable FP and those obtained from the application of noninvasive, immunochemical tests. The results from previous studies had indil ShorItlSl eyels -/8 2. First mucus 3. First t."ils "". mucus
it
.,c~ :::I
CT
:
.~
S:::I e:::I
50
20 10
(J
4. SST stlifl
'* 3 '* 4
5. PtHIk mucus
5 I
-15
-10 -5 0 .. 5 Days relative to LH peak
+10
+15
Figure 2 The cumulative frequency (%) of days on which five signals used in NFP occurred relative to the day of the LH peak.
World Health Organization Indices of the fertile period
651
99r-------1. OefintKI rise in
95 90
E,-3-G
2. OtJfined rise in ratio
20
3. Peak E,-3-G .. 4 4. PrHIk ratio .. 5
10
5 2 .•/ -10
-5
0
+5
+10
+15
Days relative to LH peak
Figure 3 The cumulative frequency (%) of days on which signals from two immunochemical tests occurred relative to the day of the LHpeak.
cated that there is a good correlation between the subjective signals used in NFP and changes in the concentration of ovarian and pituitary hormones in peripheral venous plasma or 24-hour collections ofurine. 2o- 24 Accordingly, the present study was designed to assess the potential usefulness of two immunochemical tests that might be further simplified for use in the outpatient clinic or in the home. 5 The day of the peak value for LH in EMU was used as a reference point for ovulation. With this approach there is a minimum error of ± 12 hours in the estimate, and in practice it may be ± 24 hours because of the variable time of collection for EMU and the interassay precision (coefficient of variation - 15%). The time of the probable FP (the day ofLH peak - 3 to day + 2) was calculated from the results of a study on conception risks, l which indicated that about 90% of pregnancies result from intercourse within this period. The percentage of apparent failures by each index and combined method for the start and finish of the FP must therefore be assessed in conjugation with the limitations of the reference point and the definition of a probable FP. Another problem is the units used to describe time intervals between the variables and the reference point and the duration of the possible FP. Most women and clinical investigators think in terms of calendar days, whereas the signal from each variable is obtained at a different time during the day and often repeated after a 24-hour interval. For example, an EMU is usually collected around 7:00 A.M. and 652
World Health Organization Indices of the fertile period
represents the integral value of metabolic processes occurring over the previous 8 hours. The BBT is taken on awakening and represents the temperature at an instant in time. The eM is assessed at the vulva at intervals throughout the day, and a decision is reached before retiring to bed at night. For the purpose of this study, we have defined the start of the FP as the day on which the signal was obtained, although in practice a woman could only use the information from the actual time ~f the result. Similarly, the last day of the FP was defined as the whole day on which the signal was obtained (for calculating the duration of potential fertility in days), although the time of the signal was taken as the start of the infertile period (for assessing the number of possible method failures). The use of a more precise reference point for ovulation, e.g., real-time pelvic ultrasonography, would enable the probable FP to be defined more accurately and the time intervals between variables to be described in hours. Five indices for the start of the probable FP were compared. All provided a signal in each cycle and the mean times (from - 7.0 to - 4.4) are shown in Figure 1. The smallest spread of values occurred with the day of the defined rise in E r 3-G. The proportion of signals occurring within the probable FP are shown in Figures 2 and 3 99~---------------Y-'''~~-'
95
90 80
f
1
5Q
1. SyrrplOfhermol
2. CervicallTVCus 3. BST oIcolculation 20
4. Chemical test f 5. Chemical lest 2
10
5 I
o
4
12
16
20
24
Length of apparent FP (days) Figure 4 The cumulative frequency (%) of the duration of a probable fertile period as determined by five methods.
Fertility and Sterility
Table 3. The Intra- and Interwoman Variation (One-Way Analysis of Variance) for Five Methods of Estimating the FP and the Percentage of 58 Menstrual Cycles Where the Defined FP (Day of LH Peak - 3 to +2) Is Entirely Covered Method
Symptothermal CM Calculation and BBT Chemical test 1 concentration of El-3-G Chemical test 2 ratio of E l -3-G to Pd-3a-G
Mean
SD
Minimum
Maximum
Difference between women
Withinwoman SD
% Defined FPs
covered entirely
13.4 11.9 11.8 9.3
2.9 2.9 3.3 2.2
9 5 6 6
21 19 21 15
0.2 0.04 0.03 0.1
2.7 2.6 2.9 2.0
98 91 90 83
10.7
2.3
8
17
0.001
1.9
84
and range from 1 (2%) for the ratio of E l -3-G to Pd-3a-G to 14 (24%) for the first day of fertiletype mucus. Obviously, waiting for the presence offertile-type mucus would be an unreliable indicator for family planning by periodic abstinence, and NFP programs do not suggest using the index for this purpose. In each woman there was a good agreement between the five different markers and the presumed time of ovulation. Four indices were compared for their ability to identitY the last day of potential fertility. In this connection, it should be emphasized that the peak days for E l -3-G and the ratio of E l -3-G to Pd-3a-G were identified prospectively by the algorithm as they would be in a home test and not retrospectively by eye. Once again, a sighal was obtained for all variables during each cycle. The mean values are shown in Figure 1, and it may be seen that the smallest spread of days was obtained with the day of the peak ratio of E l -3-G to Pd-3a-G + 5. The proportion of cycles where each signal fell within the probable FP is shown in Figures 2 and 3 and ranged from 3 (5%) for the day of peak mucus + 4, to 8 (14%) for the day of peak ratio + 5. Since in practice the day of peak mucus + 4 and the day of BBT rise + 3 accurately defines the end ofthe FP,13-15 we believe that the number of apparent failures (usually by 1 day) reflects the limitations of using urinary LH in EMP as a reference point for the time of ovulation. Various combinations of the start and end signals were assessed for their ability to cover the probable FP for each menstrual cycle. The symptothermal method gave the highest mean value for the apparent duration of the FP (13.4 days; range, 9 to 21 days) and entirely covered tpe whole of the probable FP in all but 2% of cycles. An examination of the data showed that the calendar calculation (shortest cycle length - 18) for the start of the FP was used in 32 cycles (55%) Vol. 39, No.5, May 1983
(P)
and the first day of mucus in 15 cycles (26%), and in 11 cycles (19%) the signal from both variables occurred on the same day. In contrast, the day of peak mucus + 4 was used to signal the last day of the FP in 28 cycles (48%) and the BBT +3 in 16 cycles (28%), and in 14 cycles (24%) the signals occurred on the same day. The mean duration of the FP delineated from the observations of eM was only slightly shorter (11.9 days), but this was associated with a corresponding increase in the number of FPs that were not completely covered (9%). The method that used the calendar calculation to identify the start of the FP and BBT rise + 3 to determine the end gave a mean duration of 11.8 days, and the number of cases where the defined FP was incompletely covered increased to 10% (mainly caused by the end signal occurring during the probable FP). At the present time, electronic devices are being produced to determine the FP by this method. 25 The immunochemical test based on concentration changes ofE l -3-G in EMU (day of defined rise to peak day + 4) gave a mean duration of 9.3 days, whereas the test based on the ratio of metabolites (day of defined rise to peak day + 5) was slightly longer (10.7 days). These methods had a similar proportion of possible failures (17% and 16%, respectively). It is of interest that the difference between women was significant for all variables other than the calculated day for the start of the FP, the rise and peak days for E l -3-G, and the peak day for the ratio of E r 3-G to Pd-3a-G. The reasons for these findings are unclear but may be related to the subjective recording of mucus and BBT by each volunteer. The significant difference between women in the day of the rise in the ratio of the metabolites may be related to whether or not an individual woman had residual activity from the corpus luteum of the previous ovarian cycle. It is interesting that the first appearance of mucus World Health Organization Indices of the fertile period
653
and the rise in the ratio of steroid glucuronides are closely associated (Table 2), both in mean interval to the LH peak and in the between- and within-woman differences. Conversely, the first day of fertile-type mucus corresponds more closely with the day of defined rise in E l -3-G. It must be recalled that all the data reported from this study were obtained from 13 welltrained, experienced users of the symptothermal method for NFP. Nevertheless, the following conclusions may be reached: (1) Both the symptothermal andCM methods of NFP best circumscribe a defined period of maximum fertility but give the longest estimates of duration. (2) A simple biophysical test using an electronic thermometer and a calculation based on cycle length may be used as an objective alternative that minimizes the amount of effort involved on the woman's part, but 10% of cycles gave signals that occurred in the probable FP. (3) Immunochemical tests based on the analysis of EMU may be used as alternatives if combined with an electronic device to record and process the daily signal. It would appear that each approach may appeal to different consumers who wish to practice family planning by periodic abstinence or restrict the use of barrier methods. In particular, the motivation and the degree of effort required by the women are important considerations in making the decision. In addition, the potential cost of the biophysical and, in particular, the biochemical methods may limit their use to a training phase in NFP. The predictive value of the two immunochemical tests may render them more suitable for monitoring the treatment of infertile couples and selecting the optimum day of intercourse to achieve a pregnancy. However, the actual value of the new tests currently under development for use in reproductive medicine can only be assessed by controlled clinical trials. Acknowledgments. The investigators are grateful to Ms. Georgina Kainer, the statistical clerk for the study, and Mrs. Pamela Martin for typing the manuscript.
REFERENCES 1. Royston JP: Basal body temperature, ovulation and the risk of conception with special reference to the lifetimes of sperm and egg. Biometrics 38:397,1982 2. WHO Task Force on Methods for the Determination of the Fertile Period: A prospective multicentre study of the ovulation method of natural family planning. m. Charac654
World Health Organization Indices of the fertile period
teristics of the menstrual cycle and of the fertile phase. Fertil Steril. In press 3. Kerin JF, Edmonds DK, Warnes GM, Cox LW, Seamark RF, Matthews CD, Young GB, Baird DT: Morphological and functional relations of Graafian follicle growth to ovulation in women using ultrasonic, laparoscopic and biochemical measurements. Br J Obstet Gynaecol 88:81, 1981 4. Moghissi KS, Syner FN, Evans TN: A composite picture of the menstrual cycle. Am J Obstet Gynecol 114:405, 1972 5. WHO Task Force on Methods for the Determination of the Fertile Period: The measurement of urinary steroid glucuronides as indices of the fertile period in women. J Steroid Biochem 17:695, 1982 6. Marshall J: The Infertile Period. London, Darton, Longman and Todd, 1973, p 93 7. Seibel MM, Smith DM, Levesque L, Borten M, Taymor ML: The temporal relationship between the luteinizing hormone surge and human oocyte maturation. Am J Obstet Gynecol 142:568, 1982 8. WHO Task Force on Methods for the Determination of the Fertile Period: Temporal relationships between ovulation and defined changes in the concentrations of plasma estradiol-17P, luteinizing hormone, follicle-stimulating hormone and progesterone. I. Probit analysis. Am J Obstet Gynecol 138:383, 1980 9. WHO Task Force on Methods for the Determination of the Fertile Period: Temporal relationships between ovulation and defined changes in the concentrations of plasma estradiol-17P, luteinizing hormone, follicle-stimulating hormone and progesterone. II. Histological dating. Am J Obstet Gynecol 139:886, 1981 10. Collins WP, Branch CM, Collins PO, Sallam HN: Biochemical indices of the fertile period in women. Int J Fertil 26:196, 1981 11. Fleming R, Coutts JRT: Prediction of ovulation in women using a rapid progesterone radioimmunoassay. Clin Endocrinol (Oxf) 16:171, 1982 12. Spieler J: Self detection of ovulation and the fertile period. In Research on Fertility and Sterility, Edited by J Cortes Prieto, A Campos da Paz, M Neves-e-Castro. Lancaster, MTP Press Ltd., 1981, p 35 13. Marshall J: A field trial of the basal-body temperature method of regulating births. Lancet 2:8, 1968 14. WHO Task Force on Methods for the Determination of the Fertile Period: A prospective multicentre trial of the ovulation method of natural family planning. I. The teaching phase. Fertil Steril 36:152, 1981 15. WHO Task Force on Methods for the Determination of the Fertile Period: A prospective multicentre trial of the ovulation method of natural family planning. II. The effectiveness phase. Fertil Steril 36:591, 1981 16. Medina JE, Cifuentes A, Abernathy JR, Spieler JM, Wade ME: Comparative evaluation of two methods of natural family planning in Colombia. Am J Obstet Gynecol 138:1142, 1980 17. Collins WP, Collins PO, Kilpatrick MJ, Manning PA, Pike J, Tyler JPP: The concentrations of oestrone3-glucuronide, LH and pregnanediol-3Ot-glucuronide as indices of ovarian function. Acta Endocrinol (Copenh) 93:123, 1979 18. Royston JP, Abrams RM: An objective method for detecting the shift in basal body temperature in women. Biometrics 36:217,1980
Fertility and Sterility
19. Royston JP, Abrams RM, Higgins MP, Flynn AM: The adjustment of basal body temperature measurements to allow for time of waking. Br J Obstet Gynaecol 87:1123, 1980 20. Billings EL, Billings JJ, Brown JB, Burger HG: Symptoms and hormonal changes accompanying ovulation. Lancet 1:282, 1972 21. Flynn AM, Lynch SS: Cervical mucus and identification of the fertile phase of the menstrual cycle. Br J Obstet Gynaecol 83:656, 1976 22. Hilgers TW, Abraham GE, Cavanagh D: Natural family planning. I. The peak symptom and estimated time of ovulation. Obstet Gynecol 52:575, 1978
Vol. 39, No.5, May 1983
23. Brown JB, Harrison P, Smith MA, Burger HG: Correlations between the mucus symptoms and the hormonal markers of fertility throughout reproductive life. Melbourne, Advocate Press Pty, 1981 24. Cortesi S, Rigoni G, Zen F, Sposetti R: Correlation of plasma gonadotrophins and ovarian steroids pattern with symptomatic changes in cervical mucus during the menstrual cycle in normal cycling women. Contraception 23:629, 1981 25. World Health Organization: 10th Annual Report, Special Programme of Research, Development and Research Training in Human Reproduction, 1981
World Health Organization Indices of the fertile period
655
FERTILITY AND STERILITY Copyright 0 1983 The American Fertility Society
Temporal relationships between indices of the fertile period
World Health Organization Task Force on Methods for the Determination of the Fertile Period, Special Programme of Research, Development and Research Training in Human Reproduction. World Health Organization, Geneva, Switzerland
The intra- and interwoman variation in nine physiologic or biochemical indices of the fertile period has been studied over 58 menstrual cycles from 13 experienced users of the symptothermal method of family planning by periodic abstinence. The time and duration of a possible fertile period have been determined by five methods (symptothermal, cervical mucus, basal body temperature plus calendar calculation, defined changes in the concentration of estrone-3 -glucuronide and the ratio of estrone-3-glucuronide to pregnanediol-3a-glucuronide as determined by immunochemical tests on daily samples of early morning urine). The values were compared with a period of probable fertility (day of urinary luteotropin [LHJ peak - 3 to day of LH peak + 2). The duration of the possible fertile period by each method (mean ± standard deviation) was 13.4 (2.9), 11.9 (2.9), 11.8 (3.3),9.3 (2.2), and 10.9 (2.3) days, respectively, while the percentage of the probable fertile periods covered entirely by each approach was 98%, 91%, 90%, 83%, and 84%, respectively. The results warrant the initiation of clinical trials to ascertain the practical value of the individual or combined tests for family planning and the management of infertility. Fertil Steril 39:647, 1983
The true fertile period (FP) in women only occurs when there is a viable ovum. Ovulation is the crucial event during the menstrual cycle, and it has been calculated that the maximum survival time for the ovum is probably less than 48 hours. 1 Received September 13, 1982; revised and accepted December 22, 1982. Reprint requests: Professor W. P. Collins, Department of Obstetrics and Gynaecology, King's College Hospital, Denmark Hill, London SE5 8RX, United Kingdom. The Task Force Investigators for this study were William Collins, Ph.D., D.Sc., Janet Simpson, B.Sc., and Patricia Collins, B.Sc., Department of Obstetrics and Gynaecology, King's College Hospital, London, United Kingdom; Anna Flynn, M.R.C.O.G., and Maureen McGarry, Birmingham Maternity Hospital, Queen Elizabeth Medical Centre, Birmingham, United Kingdom; Patrick Royston, B.A., M.Sc., M.I.S., Division of Computing and Statistics, MRC Clinical Research Centre, Harrow, United Kingdom; and Jeffrey Spieler, M.Sc., and Alain Pinol, D.E.S.T., Data and Text Processing Services Unit, Division ofInformation Systems Support, World Health Organization, Geneva, Switzerland. Vol. 39, No.5, May 1983
After intercourse, the spermatozoa may have a longer life span (up to 6 days), and their fertilizing potential is dependent upon the quality of mucus in the female genital tract, which enables the gametes to meet in one of the oviducts. Hence, the potentially FP of a couple is the time from the first act of intercourse which may lead to pregnancy to the demise of the ovum. Accordingly, attempts have been made to identify three phases during each menstrual cycle: (1) an initial phase when there is only a low risk that intercourse could lead to conception; (2) a periovulatory phase when the chance of fertilization increases to a maximum and then rapidly falls; and (3) a phase of absolute infertility when the ovum is no longer viable. 1 In practice, there may be more than one apparent FP each menstrual cycle. 2 The application of radioimmunoassay (RIA) and pelvic ultrasonography to the daily monitoring of healthy women has established that the development of an ovarian follicle is associated World Health Organization Indices of the fertile period
647
with an increase in the concentration of estradiol (E 2) in peripheral venous plasma. 3 Furthermore, the rapidly increasing levels of this hormone in the peripheral circulation stimulates an increase in the production and changes in the characteristics of cervical mucus (CM) and subsequently initiates a surge in the release ofluteotropin (LH) from the pituitary.4 Consequently, defined changes in the concentration of E2 in serial samples of peripheral plasma or daily changes in the characteristics of CM have been used to indicate ovulation and the probable start of the FP. Other studies with labeled E2 and RIAs have shown that from 50% to 80% of the circulating hormone is excreted in the urine over a period of 24 hours and that estrone-3-glucuronide (E 1-3-G) is one of the principal metabolites. The concentration of this conjugate may be measured in diluted samples of early morning urine (EMU) without prior hydrolysis or extraction and characteristic changes used to estimate the start of the probable FP. 5 Alternatively, the likely start of each FP has been calculated from a knowledge of the previous menstrual history, e.g., the shortest length ofthe previous six cycles minus 18 days. 6 The resumption of oogenesis, 7 follicular rupture, and luteinization of the theca and granulosa cellss, 9 occurs after a noticeable increase in the level of circulating LH. For example, the results from a World Health Organization (WHO) multicenter studys indicated that a defined rise in the level of plasma LH (a 50% increase in the mean of three preceding values; sampling every 8 hours) was a good index of impending ovulation, which occurred from 24 to 56 hours later (mean, 32 hours). Moreover, LH is rapidly excreted in urine, and the time interval from the rise of the metabolite in EMU to follicular rupture as determined by ultrasonography is 20 to 44 hours, with a mean of 30 hours. 1o Luteinization, in terms of increased progesterone (P) in the peripheral circulation, immediately precedes ovulations, 11 and is associated with changes in the amount and physicochemical properties of CM. Hence, an appropriate rise in the concentration of plasma P or a change in the characteristics of CM12 may be used as reference points to calculate the end of the FP. The clearance ofP from the vascular compartment (approximately 25% over 24 hours) is slower than for E2 and may explain why there are limitations to the use of changes in the concentration of pregnanediol-3a-glucuronide (Pd-3a-G) to locate the end 648
World Health Organization Indices of the fertik period
of the FP. 5 However, the observation that basal body temperature (BBT) rises around the time of follicular rupture has been used to calculate the start of the postovulatory infertile phase of the ovarian cycle. 12, 13 The present study was prompted by results on the practical value of the CM 14,15 and symptothermal methods 16 to delineate the FP and the recent development of immunochemical tests based on the analysis of steroid glucuronides in EMU to achieve the same objective. 5 The aim was to compare (both within and between women) the time relationship between five different indices of the start and four of the finish of each FP. In addition, the duration of the FP was calculated by five methods (symptothermal, CM, BBT plus calendar calculation, and immunochemical tests based either on changes in the concentration of E 1-3-G in EMU or in the derived ratio E 1-3-G to Pd-3a-G). The results are compared in terms of the accuracy with which the FP is located and the time of abstinence required if the method had been used to avoid a pregnancy.
MATERIALS AND METHODS
The intention was to study at least ten healthy women for six consecutive menstrual cycles. All potential participants were informed of the purpose of the investigation and reassured that they would be free to withdraw at any time. The subjects were recruited from the Natural Family Planning Clinic at Birmingham Maternity Hospital, and physiologic indices of the FP were recorded in the home on a special form provided by WHO. Daily samples of EMU were collected, stored at - 15° C, and subsequently sent to King's College Hospital Medical School, London, for the analysis ofE 1-3-G, LH, and Pd-3a-G. The results were also recorded on forms provided by WHO. All forms (from Birmingham and London) were sent independently to WHO in Geneva for data processing. SUBJECTS
Thirteen experienced users of the symptothermal method of family planning were studied. They were between 25 and 33 years of age, and ten had had at least one child. For inclusion in the study, each woman had to have experienced regular menstrual cycles (25 to 35 days for six previous cycles) and not be receiving any form of Fertility and Sterility
chronic drug therapy or working night shifts. In addition, no subject had a history of or evidence of liver or kidney disease or dysfunction or had used any form of hormonal contraception or intrauterine device during the previous three menstrual cycles. All participants had to be willing to complete symptothermal charts· and collect EMU for six complete menstrual cycles. METHODS
The BBT was taken each morning, before getting out of bed. A mercury thermometer with an expanded scale was used. The presence and type of CM at the vulva was assessed according to sensation and described as dry, damp, or wet. In addition, each subject wiped the vulva with toilet paper before and after micturition, and recorded the sensation and appearance of the mucus. Observations were made throughout the day and an assessment was made immediately before retiring to bed at night. The concentrations of E l -3-G, LH, and Pd-3aG were measured by RIA without prior hydrolysis or extraction. The methods have been described and evaluated in detail. 17 DEFINITIONS
The following definitions were used. Day 1 of the menstrual cycle was the first day of menstruation, i.e., the day starting after midnight on which there was a noticeable flow of blood (requiring sanitary protection) through the vagina. The day of LH peak was the day of the highest arithmetic value for urinary LH between days 5 and 25 of the menstrual cycle. The probable FP was defined as the day of the LH peak - 3 to the day of the LH peak + 2, inclusive. A signal on day LH peak - 3 was accepted as indicating the start of the FP and on day LH peak + 2 as the beginning of the infertile phase. The calculated start of the FP was the day of the current menstrual cycle resulting from the calculation (number of days in the shortest of six previous cycles -18). The length of each new cycle was subsequently used in the calculation. The first day of mucus was the day on which there was a sensation of dampness (i.e., no longer dry) or the first visual appearance of mucus. The first day of fertile-type mucus was the first day that a sensation of wetness or slippery feeling was recorded (with or without the appearance of any kind of mucus) or the appearance of clear stretchy mucus. The day of peak Vol. 39, No.5, May 1983
mucus was the last day of the fertile-type mucus, which could only be determined retrospectively, i.e., on the next day. An EMU was defined as the first urine passed upon awakening in the morning. The days on which a sustained rise occurred in the concentration of E l -3-G and E l -3-GIPd-3a-G were determined by an adaptation of the cumulative sum (CUSUM) test. 18 All values were converted to IOg10 before analysis. A baseline was calculated as the mean of values on days 1 to 6 for each cycle studied. A reference level was calculated as baseline + 1 standard deviation (SD). A single value for the mean population SD was found to be adequate for each of the variables and was 0.30 for E l -3-G and 0.35 for the ratio, both in log units. The CUSUM test was performed by subtracting the reference level from each daily value and adding the result to the current CUSUM. A significant rise was recorded when the CUSUM reached a decision level, which was taken as 2 SD above the mean baseline value. The day of the defined rise (used to indicate the start of the FP) was the day on which the data first exceeded the reference level in the run leading to a significant CUSUM. The peak days for E l -3-G and the ratio were determined by starting with the rise day + 1 and taking the day with the highest value that was followed by three consecutive lower amounts. The third day of higher phase (postovulatorypremenstrual) BBT was also identified by a modified CUSUM technique. 18 The baseline was taken as the mean temperature of 8 days, starting at day 4, and the reference level as baseline plus 0.1 0 C for each cycle. The CUSUM showed a significant rise when it exceeded 0.25; however, this could occur before three high temperatures had been recorded. For standardization of the procedure, three consecutive temperatures greater than or equal to the reference value were required before the third of these was selected as the last day of the FP (BBT + 3). Days were excluded on which the subject recorded more than 1 day of illness, which elevated the temperature. If the time of taking the BBT was recorded as later than usual, the reading was reduced by 0.10 C for each additional hour.19 LIMITS OF THE FERTILE PERIOD
All indicators of the FP were timed relative to the day of the LH peak; for example, day -1 was World Health Organization Indices of the fertile period
649
Table 1. The Number of Subjects and Cycles Studied
Total
No. of subjects
No. of menstrual cycles/subject
7 1 2 2 1
6 5 4 2 1
No. of conceptional cycles a
13
(
No. of cycles excluded from analysisb
42 4 7 4 1
1 1
1 1
2
2
Total no. of cycles analyzed
58
alncluded for data processing. bExcluded because of missing samples of EMU.
the day preceding the peak, and day + 1 was 1 day after the peak. There was five potential markers of the first fertile day: F1, the shortest of the last six menstrual cycle lengths - 18; F2, the first day of any eM; F3, the first day of fertiletype eM; F4, the day of defined rise in E1"3-G; and F5, the day of a defined rise in E l -3-G/Pd3a-G. Four indices of the end of the FP were compared: L1, the day of peak E1"3-G + 4; L2, the day of peak mucus + 4; L3, the third day of high BBT (BBT + 3); and L4, the day of peak ratio +
peak + 2, (i.e., the probable FP). One-way analysis of variance was used to determine whether there were significant differences between women for any of the variables. Similar analyses were applied to the five methods of calculating the duration of the FP (D1 to D5). In particular, the proportion of cycles was calculated in which each method failed to cover the period of the day of LH peak -3 to day ofLH peak +2.
RESULTS
5. LENGTH OF THE FERTILE PERIOD
The duration of the FP was estimated by five methods: D1, symptothermal (later of L2 and L3 - earlier of F1 and F2 + 1); D2, eM (L2 - F2 + 1); D3, calculation and BBT (L3 - F2 + 1); D4, chemical test 1-E l -3-G (L1 - F4 + 1); and D5, chemical test 2-ratio of E l -3-G to Pd-3a-G (L5 - F5 + 1). D1, D2, and D3 represent variants of techniques used in natural family planning (NFP) programs, and D4 and D5 are immunochemical tests that might be considered as alternatives to D1, D2, or D3. STATISTICAL ANALYSIS
The purpose of the analysis was to examine the time intervals between five indices to predict the start of the FP (F1 to F5) and the day ofLH peak, and between four indices to detect the end (L1 to L4) from the same reference point. A signal was said to have occurred on a given day if the start or end of the fertile period was indicated by one of the methods. The cumulative proportion of cycles with each signal was plotted on probability paper against the day of the signal relative to the day of the LH peak. In this way, the number of failures and successes of each index was determined by reference to the day of LH peak - 3 to day of LH 650
World Health Organization Indices of the fertile period
The number of subjects recruited and menstrual cycles analyzed are shown in Table 1. The results from two cycles were excluded from the analysis because samples of EMU were missing for at least 5 consecutive days during the time of follicular development. An additional 40 samples of EMU were missing « 3% of total), but this limitation would not have affected the temporal relationship of any hormonal index by more than 24 hours.
Peak mucus +4 BBTshift +3
--. Mean and
observed n = 58
Peak ratio +5 Peak E.-3-G +4
First fertil. mucus Defined rise in Defined riee In First mucus Short.st cycle -18
•
-10
Figure 1
-5 .0 +5 +10 Days relative to LH peak
The means and 90% observed ranges for the days of nine signals relative to the day of the LH peak.
Fertility and Sterility
Table 2. Intra- and Interwoman Variation-One-Way Analysis of Variance for Nine Indices of the Start or Finish of the FP Relative to the Day of LH Peaka Variable Length of menstrual cycle Day of LH peak Shortest cycle length -18 First day of mucus First day of fertiletype mucus Day of defined rise in E l -3-G Day of defined rise in ratio Day of peak E l -3-G +4 Day of peak mucus +4 BBT rise +3 Day of peak ratio +5
Mean
SD
Minimum
Difference between women
Maximum
(P)
Witbinwoman SD
No. of cycles with signal withinFP
27·.7
2.7
22
35
0.0001
2.0
15.0 -7.0
2.6 2.5
10 -14
23 -1
0.0001 0.3
2.0 2.5
2
-6.6 -4.4
2.5 2.6
-12
-11
-2 0
0.0002 0.03
2.0 2.3
1 14
-4.9
1.9
-10
-1
0.3
1.8
5
-6.5
2.8
-14
-1
0.007
2.4
1
3.4
1.9
-4
9
0.09
1.7
5
4.3
2.2
-5
10
0.055
2.0
3
3.8 3.2
2.0 2.3
-1 -4
11
0.0008 0.7
1.6 2.3
4 8
6
aOverall means and ranges for 58 menstrual cycles and number of inadequate signals.
INDICES FOR THE LIMITS OF THE FERTILE PERIOD
The means and 90% observed limits for the nine indices are shown in Figure 1 relative to the probable period of fertility. The 90% limits were calculated as the 4th and 55th ordered values in the sample of 58 cycles. The numeric values for the means, SDs, and ranges of the indices, together with the lengths of the menstrual cycles and the days of the LH peaks, are listed in Table 2. The significance of difference for each variable between women, together with the SD within women, is also shown in Table 2. The withinwoman SD represents the variability of values from cycle to cycle in each subject. The cumulative percentage of cycles giving each signal used in NFP on days relative to the day of the LH peak is shown in Figure 2; and from those used in the immunochemical tests, in Figure 3. The number of cycles where each signal falls within the defined FP (Le., late indication of the start or falsepositive detection end of the FP) is listed in Table
2. DURATION OF THE FERTILE PERIOD
The apparent duration of the FP for all cycles as determined by the five combined indices (D1 to D5) is shown as frequency distribution curves in Figure 4. The means, SDs, and ranges of values as determined by each method, together with the significance of difference between intra- and inVol. 39, No.5, May 1983
terwoman variation, are shown in Table 3. The number of cycles where the calculated FP by each method did not entirely cover the defined FP ranged between 2% and 17%. DISCUSSION
This report describes a prospective study of the time intervals between indices used in NFP to locate the probable FP and those obtained from the application of noninvasive, immunochemical tests. The results from previous studies had indil ShorItlSl eyels -/8 2. First mucus 3. First t."ils "". mucus
it
.,c~ :::I
CT
:
.~
S:::I e:::I
50
20 10
(J
4. SST stlifl
'* 3 '* 4
5. PtHIk mucus
5 I
-15
-10 -5 0 .. 5 Days relative to LH peak
+10
+15
Figure 2 The cumulative frequency (%) of days on which five signals used in NFP occurred relative to the day of the LH peak.
World Health Organization Indices of the fertile period
651
99r-------1. OefintKI rise in
95 90
E,-3-G
2. OtJfined rise in ratio
20
3. Peak E,-3-G .. 4 4. PrHIk ratio .. 5
10
5 2 .•/ -10
-5
0
+5
+10
+15
Days relative to LH peak
Figure 3 The cumulative frequency (%) of days on which signals from two immunochemical tests occurred relative to the day of the LHpeak.
cated that there is a good correlation between the subjective signals used in NFP and changes in the concentration of ovarian and pituitary hormones in peripheral venous plasma or 24-hour collections ofurine. 2o- 24 Accordingly, the present study was designed to assess the potential usefulness of two immunochemical tests that might be further simplified for use in the outpatient clinic or in the home. 5 The day of the peak value for LH in EMU was used as a reference point for ovulation. With this approach there is a minimum error of ± 12 hours in the estimate, and in practice it may be ± 24 hours because of the variable time of collection for EMU and the interassay precision (coefficient of variation - 15%). The time of the probable FP (the day ofLH peak - 3 to day + 2) was calculated from the results of a study on conception risks, l which indicated that about 90% of pregnancies result from intercourse within this period. The percentage of apparent failures by each index and combined method for the start and finish of the FP must therefore be assessed in conjugation with the limitations of the reference point and the definition of a probable FP. Another problem is the units used to describe time intervals between the variables and the reference point and the duration of the possible FP. Most women and clinical investigators think in terms of calendar days, whereas the signal from each variable is obtained at a different time during the day and often repeated after a 24-hour interval. For example, an EMU is usually collected around 7:00 A.M. and 652
World Health Organization Indices of the fertile period
represents the integral value of metabolic processes occurring over the previous 8 hours. The BBT is taken on awakening and represents the temperature at an instant in time. The eM is assessed at the vulva at intervals throughout the day, and a decision is reached before retiring to bed at night. For the purpose of this study, we have defined the start of the FP as the day on which the signal was obtained, although in practice a woman could only use the information from the actual time ~f the result. Similarly, the last day of the FP was defined as the whole day on which the signal was obtained (for calculating the duration of potential fertility in days), although the time of the signal was taken as the start of the infertile period (for assessing the number of possible method failures). The use of a more precise reference point for ovulation, e.g., real-time pelvic ultrasonography, would enable the probable FP to be defined more accurately and the time intervals between variables to be described in hours. Five indices for the start of the probable FP were compared. All provided a signal in each cycle and the mean times (from - 7.0 to - 4.4) are shown in Figure 1. The smallest spread of values occurred with the day of the defined rise in E r 3-G. The proportion of signals occurring within the probable FP are shown in Figures 2 and 3 99~---------------Y-'''~~-'
95
90 80
f
1
5Q
1. SyrrplOfhermol
2. CervicallTVCus 3. BST oIcolculation 20
4. Chemical test f 5. Chemical lest 2
10
5 I
o
4
12
16
20
24
Length of apparent FP (days) Figure 4 The cumulative frequency (%) of the duration of a probable fertile period as determined by five methods.
Fertility and Sterility
Table 3. The Intra- and Interwoman Variation (One-Way Analysis of Variance) for Five Methods of Estimating the FP and the Percentage of 58 Menstrual Cycles Where the Defined FP (Day of LH Peak - 3 to +2) Is Entirely Covered Method
Symptothermal CM Calculation and BBT Chemical test 1 concentration of El-3-G Chemical test 2 ratio of E l -3-G to Pd-3a-G
Mean
SD
Minimum
Maximum
Difference between women
Withinwoman SD
% Defined FPs
covered entirely
13.4 11.9 11.8 9.3
2.9 2.9 3.3 2.2
9 5 6 6
21 19 21 15
0.2 0.04 0.03 0.1
2.7 2.6 2.9 2.0
98 91 90 83
10.7
2.3
8
17
0.001
1.9
84
and range from 1 (2%) for the ratio of E l -3-G to Pd-3a-G to 14 (24%) for the first day of fertiletype mucus. Obviously, waiting for the presence offertile-type mucus would be an unreliable indicator for family planning by periodic abstinence, and NFP programs do not suggest using the index for this purpose. In each woman there was a good agreement between the five different markers and the presumed time of ovulation. Four indices were compared for their ability to identitY the last day of potential fertility. In this connection, it should be emphasized that the peak days for E l -3-G and the ratio of E l -3-G to Pd-3a-G were identified prospectively by the algorithm as they would be in a home test and not retrospectively by eye. Once again, a sighal was obtained for all variables during each cycle. The mean values are shown in Figure 1, and it may be seen that the smallest spread of days was obtained with the day of the peak ratio of E l -3-G to Pd-3a-G + 5. The proportion of cycles where each signal fell within the probable FP is shown in Figures 2 and 3 and ranged from 3 (5%) for the day of peak mucus + 4, to 8 (14%) for the day of peak ratio + 5. Since in practice the day of peak mucus + 4 and the day of BBT rise + 3 accurately defines the end ofthe FP,13-15 we believe that the number of apparent failures (usually by 1 day) reflects the limitations of using urinary LH in EMP as a reference point for the time of ovulation. Various combinations of the start and end signals were assessed for their ability to cover the probable FP for each menstrual cycle. The symptothermal method gave the highest mean value for the apparent duration of the FP (13.4 days; range, 9 to 21 days) and entirely covered tpe whole of the probable FP in all but 2% of cycles. An examination of the data showed that the calendar calculation (shortest cycle length - 18) for the start of the FP was used in 32 cycles (55%) Vol. 39, No.5, May 1983
(P)
and the first day of mucus in 15 cycles (26%), and in 11 cycles (19%) the signal from both variables occurred on the same day. In contrast, the day of peak mucus + 4 was used to signal the last day of the FP in 28 cycles (48%) and the BBT +3 in 16 cycles (28%), and in 14 cycles (24%) the signals occurred on the same day. The mean duration of the FP delineated from the observations of eM was only slightly shorter (11.9 days), but this was associated with a corresponding increase in the number of FPs that were not completely covered (9%). The method that used the calendar calculation to identify the start of the FP and BBT rise + 3 to determine the end gave a mean duration of 11.8 days, and the number of cases where the defined FP was incompletely covered increased to 10% (mainly caused by the end signal occurring during the probable FP). At the present time, electronic devices are being produced to determine the FP by this method. 25 The immunochemical test based on concentration changes ofE l -3-G in EMU (day of defined rise to peak day + 4) gave a mean duration of 9.3 days, whereas the test based on the ratio of metabolites (day of defined rise to peak day + 5) was slightly longer (10.7 days). These methods had a similar proportion of possible failures (17% and 16%, respectively). It is of interest that the difference between women was significant for all variables other than the calculated day for the start of the FP, the rise and peak days for E l -3-G, and the peak day for the ratio of E r 3-G to Pd-3a-G. The reasons for these findings are unclear but may be related to the subjective recording of mucus and BBT by each volunteer. The significant difference between women in the day of the rise in the ratio of the metabolites may be related to whether or not an individual woman had residual activity from the corpus luteum of the previous ovarian cycle. It is interesting that the first appearance of mucus World Health Organization Indices of the fertile period
653
and the rise in the ratio of steroid glucuronides are closely associated (Table 2), both in mean interval to the LH peak and in the between- and within-woman differences. Conversely, the first day of fertile-type mucus corresponds more closely with the day of defined rise in E l -3-G. It must be recalled that all the data reported from this study were obtained from 13 welltrained, experienced users of the symptothermal method for NFP. Nevertheless, the following conclusions may be reached: (1) Both the symptothermal andCM methods of NFP best circumscribe a defined period of maximum fertility but give the longest estimates of duration. (2) A simple biophysical test using an electronic thermometer and a calculation based on cycle length may be used as an objective alternative that minimizes the amount of effort involved on the woman's part, but 10% of cycles gave signals that occurred in the probable FP. (3) Immunochemical tests based on the analysis of EMU may be used as alternatives if combined with an electronic device to record and process the daily signal. It would appear that each approach may appeal to different consumers who wish to practice family planning by periodic abstinence or restrict the use of barrier methods. In particular, the motivation and the degree of effort required by the women are important considerations in making the decision. In addition, the potential cost of the biophysical and, in particular, the biochemical methods may limit their use to a training phase in NFP. The predictive value of the two immunochemical tests may render them more suitable for monitoring the treatment of infertile couples and selecting the optimum day of intercourse to achieve a pregnancy. However, the actual value of the new tests currently under development for use in reproductive medicine can only be assessed by controlled clinical trials. Acknowledgments. The investigators are grateful to Ms. Georgina Kainer, the statistical clerk for the study, and Mrs. Pamela Martin for typing the manuscript.
REFERENCES 1. Royston JP: Basal body temperature, ovulation and the risk of conception with special reference to the lifetimes of sperm and egg. Biometrics 38:397,1982 2. WHO Task Force on Methods for the Determination of the Fertile Period: A prospective multicentre study of the ovulation method of natural family planning. m. Charac654
World Health Organization Indices of the fertile period
teristics of the menstrual cycle and of the fertile phase. Fertil Steril. In press 3. Kerin JF, Edmonds DK, Warnes GM, Cox LW, Seamark RF, Matthews CD, Young GB, Baird DT: Morphological and functional relations of Graafian follicle growth to ovulation in women using ultrasonic, laparoscopic and biochemical measurements. Br J Obstet Gynaecol 88:81, 1981 4. Moghissi KS, Syner FN, Evans TN: A composite picture of the menstrual cycle. Am J Obstet Gynecol 114:405, 1972 5. WHO Task Force on Methods for the Determination of the Fertile Period: The measurement of urinary steroid glucuronides as indices of the fertile period in women. J Steroid Biochem 17:695, 1982 6. Marshall J: The Infertile Period. London, Darton, Longman and Todd, 1973, p 93 7. Seibel MM, Smith DM, Levesque L, Borten M, Taymor ML: The temporal relationship between the luteinizing hormone surge and human oocyte maturation. Am J Obstet Gynecol 142:568, 1982 8. WHO Task Force on Methods for the Determination of the Fertile Period: Temporal relationships between ovulation and defined changes in the concentrations of plasma estradiol-17P, luteinizing hormone, follicle-stimulating hormone and progesterone. I. Probit analysis. Am J Obstet Gynecol 138:383, 1980 9. WHO Task Force on Methods for the Determination of the Fertile Period: Temporal relationships between ovulation and defined changes in the concentrations of plasma estradiol-17P, luteinizing hormone, follicle-stimulating hormone and progesterone. II. Histological dating. Am J Obstet Gynecol 139:886, 1981 10. Collins WP, Branch CM, Collins PO, Sallam HN: Biochemical indices of the fertile period in women. Int J Fertil 26:196, 1981 11. Fleming R, Coutts JRT: Prediction of ovulation in women using a rapid progesterone radioimmunoassay. Clin Endocrinol (Oxf) 16:171, 1982 12. Spieler J: Self detection of ovulation and the fertile period. In Research on Fertility and Sterility, Edited by J Cortes Prieto, A Campos da Paz, M Neves-e-Castro. Lancaster, MTP Press Ltd., 1981, p 35 13. Marshall J: A field trial of the basal-body temperature method of regulating births. Lancet 2:8, 1968 14. WHO Task Force on Methods for the Determination of the Fertile Period: A prospective multicentre trial of the ovulation method of natural family planning. I. The teaching phase. Fertil Steril 36:152, 1981 15. WHO Task Force on Methods for the Determination of the Fertile Period: A prospective multicentre trial of the ovulation method of natural family planning. II. The effectiveness phase. Fertil Steril 36:591, 1981 16. Medina JE, Cifuentes A, Abernathy JR, Spieler JM, Wade ME: Comparative evaluation of two methods of natural family planning in Colombia. Am J Obstet Gynecol 138:1142, 1980 17. Collins WP, Collins PO, Kilpatrick MJ, Manning PA, Pike J, Tyler JPP: The concentrations of oestrone3-glucuronide, LH and pregnanediol-3Ot-glucuronide as indices of ovarian function. Acta Endocrinol (Copenh) 93:123, 1979 18. Royston JP, Abrams RM: An objective method for detecting the shift in basal body temperature in women. Biometrics 36:217,1980
Fertility and Sterility
19. Royston JP, Abrams RM, Higgins MP, Flynn AM: The adjustment of basal body temperature measurements to allow for time of waking. Br J Obstet Gynaecol 87:1123, 1980 20. Billings EL, Billings JJ, Brown JB, Burger HG: Symptoms and hormonal changes accompanying ovulation. Lancet 1:282, 1972 21. Flynn AM, Lynch SS: Cervical mucus and identification of the fertile phase of the menstrual cycle. Br J Obstet Gynaecol 83:656, 1976 22. Hilgers TW, Abraham GE, Cavanagh D: Natural family planning. I. The peak symptom and estimated time of ovulation. Obstet Gynecol 52:575, 1978
Vol. 39, No.5, May 1983
23. Brown JB, Harrison P, Smith MA, Burger HG: Correlations between the mucus symptoms and the hormonal markers of fertility throughout reproductive life. Melbourne, Advocate Press Pty, 1981 24. Cortesi S, Rigoni G, Zen F, Sposetti R: Correlation of plasma gonadotrophins and ovarian steroids pattern with symptomatic changes in cervical mucus during the menstrual cycle in normal cycling women. Contraception 23:629, 1981 25. World Health Organization: 10th Annual Report, Special Programme of Research, Development and Research Training in Human Reproduction, 1981
World Health Organization Indices of the fertile period
655