a prototype for ovulation detection: pros and cons

5
A prototype for ovulation detection: Pros and cons Bill L. Lasley, PhD, Susan E. Shideler, PhD, and Coralie J. Munro, BSc Davis, California A noninstrumented enzyme immunoassay for urinary estrone conjugates was adapted from an instrumented microtiter plate enzyme immunoassay assay. The end point of the assay was a color change from green to clear, which was visible to the unaided eye. The visible color change was adjusted to allow 80 ng l ml estrone conjugates (on the basis of a sample size of 6.5 ,...1 urine) to be distinguished from an infinite dilution without instrumentation. The evaluation of human urine collected from ovulatory ovarian cycles demonstrated that early follicular phase concentrations (35.9 ± 6.8 to 79.4 ± 14.7 nglml, n = 10) produced a dark-green color, whereas late follicular phase concentrations (162.9 ± 20.1 nglml, n = 10) produced no color. Daily urine samples throughout 10 ovulatory ovarian cycles produced parallel profiles when compared to measurements of estradiol in paired blood samples. Complete analysis of the data indicated that ovarian follicular dynamics can be accurately monitored through the noninstrumented analysis of daily estrone conjugates in urine samples. (AM J OSSTET GVNECOL 1991 ;165:2003-7.) The events associated with ovarian follicular growth and development traditionally have been characterized by the measurement of serum or plasma estradiol. More recently, radioimmunoassay (RIA) measure- ments of urinary metabolites have been shown to be practical replacements for measurements of blood con- centrations of estradiol in human' -3 and nonhuman pri- mates.' Urinary measurements of excreted hormone concentrations have been shown to provide an accurate appraisal of estrogen dynamics during both the normal and the abnormal human menstrual cycle as well as during the periimplantation period in human be- ings 3 . 5 and macaque monkeys.' At the present time, efforts are being made to pro- vide nonradiometric assay substitutes for RIA methods . Enzyme immunoassays (EIAs) have been developed for urinary steroid metabolites. 6 8 Urinary analysis of ovar- ian function has advantages over serum analysis be- cause subjects can collect their own samples for lengthy studies and store them by freezing without preserva- tives for analysis at a central site. Enzyme assay formats have advantages over RIAs in that they eliminate toxic and regulated substances, reduce monetary and labor costs of analysis, and enhance the ability to expand study designs to include the monitoring of large num- bers of subjects over relatively large geographic areas. Additionally, enzyme immunoassay formats provide From the Department of Reproduction, School of Veterinary M edicine, University of California. Supported by The National Institute of International Studies in Nat- ural Family Planning under a cooperative agreement with the U.S. Agency for International Development DPE-3040-AOO-5064-00. Reprint requests: Bill Lasley, PhD, Department of R eproduction, School of Veterinary M edicine, Universi ty of California, Davis, CA 95616. 610134954 the basis for the development of noninstrumented as- says that can be used under nonlaboratory conditions. Such assays would allow subjects to collect and analyze their own daily urine samples and , in certain situations, detect the fertile period or anticipate ovulation. Commercial tests are currently available for the pri- mary metabolites of progesterone, but there are few, if any, tests for impending ovulation. Most tests, like those signaling the luteinizing hormone (LH) surge, or change in the estrogen / progesterone ratio, allow de- tection or confirmation of ovulation only after the fer- tile period has begun . Such indicators can be used for fertility enhancement , but they have little value as an aid to prospectively avoid conception. The best indicator of the fertile period is the initiation of the preovulatory estrogen rise. This event coincides with the selection of the ovarian follicle, which will ovu- late 3 to 6 days later. If the initial rise in serum estradiol is predictive of ovulation by 5 days, then it would follow that urinary metabolites also can be used for the same prediction-assuming a minimal lag time between changes in serum estradiol and urinary metabolites. We have demonstrated a preovulatory rise in both serum and urinary estrogen ,s but have not correlated the events in urine to those in serum using a noninstru- mented assay format that could be used to document follicular development and ovulation outside of the lab- oratory. The purpose of this study was to evaluate the ad- aptation of an instrumented EIA for estrone conjugates (EIC) to a noninstrumented EIA (NEIA). Changes in urinary EIC concentrations measured by NEIA were compared to changes in serum estradiol concentrations in paired daily blood and urine samples throughout ovulatory menstrual cycles. Comparisons were made A prototype for ovulation detection: Pros and cons Bill L. Lasley, PhD, Susan E. Shideler, PhD, and Coralie J. Munro, BSe Davis, California A noninstrumented enzyme immunoassay for urinary estrone conjugates was adapted from an instrumented microtiter plate enzyme immunoassay assay. The end point of the assay was a color change from green to clear, which was visible to the unaided eye. The visible color change was adjusted to allow 80 nglml estrone conjugates (on the basis of a sample size of 6.5 fJol urine) to be distinguished from an infinite dilution without instrumentation. The evaluation of human urine collected from ovulatory ovarian cycles demonstrated that early follicular phase concentrations (35.9 ± 6.8 to 79.4 ± 14.7 nglml, n = 10) produced a dark-green color, whereas late follicular phase concentrations (162.9 ± 20.1 nglml, n = 10) produced no color. Daily urine samples throughout 10 ovulatory ovarian cycles produced parallel profiles when compared to measurements of estradiol in paired blood samples. Complete analysis of the data indicated that ovarian follicular dynamics can be accurately monitored through the noninstrumented analysis of daily estrone conjugates in urine samples. (AM J OBSTET GVNECOL 1991;165:2003-7.) The events associated with ovarian follicular growth and development traditionally have been characterized by the measurement of serum or plasma estradiol. More recently, radioimmunoassay (RIA) measure- ments of urinary metabolites have been shown to be practical replacements for measurements of blood con- centrations of estradiol in human 1·3 and nonhuman pri- mates.' Urinary measurements of excreted hormone concentrations have been shown to provide an accurate appraisal of estrogen dynamics during both the normal and the abnormal human menstrual cycle as well as during the periimplantation period in human be- ings3.5 and macaque monkeys.' At the present time, efforts are being made to pro- vide nonradiometric assay substitutes for RIA methods. Enzyme immunoassays (EIAs) have been developed for urinary steroid metabolites. 6 s Urinary analysis of ovar- ian function has advantages over serum analysis be- cause subjects can collect their own samples for lengthy studies and store them by freezing without preserva- tives for analysis at a central site. Enzyme assay formats have advantages over RIAs in that they eliminate toxic and regulated substances, reduce monetary and labor costs of analysis, and enhance the ability to expand study designs to include the monitoring of large num- bers of subjects over relatively large geographic areas. Additionally, enzyme immunoassay formats provide From the Department of Reproduction, School ofVete1-inary Medicine, University of California. Supported by The National Institute of International Studies in Nat- ural Family Planning under a cooperative agreement with the U.S. Agency for International Development DPE-3040-AOO-5064-00. Reprint requests: Bill Lasley, PhD, Department of Reproduction, School of Veterinary Medicine, University of California, Davis, CA 95616. 610134954 the basis for the development of noninstrumented as- says that can be used under nonlaboratory conditions. Such assays would allow subjects to collect and analyze their own daily urine samples and, in certain situations, detect the fertile period or anticipate ovulation. Commercial tests are currently available for the pri- mary metabolites of progesterone, but there are few, if any, tests for impending ovulation. Most tests, like those signaling the luteinizing hormone (LH) surge, or change in the estrogen/progesterone ratio, allow de- tection or confirmation of ovulation only after the fer- tile period has begun. Such indicators can be used for fertility enhancement, but they have little value as an aid to prospectively avoid conception. The best indicator of the fertile period is the initiation of the preovulatory estrogen rise. This event coincides with the selection of the ovarian follicle, which will ovu- late 3 to 6 days later. If the initial rise in serum estradiol is predictive of ovulation by 5 days, then it would follow that urinary metabolites also can be used for the same prediction-assuming a minimal lag time between changes in serum estradiol and urinary metabolites. We have demonstrated a preovulatory rise in both serum and urinary estrogen,s but have not correlated the events in urine to those in serum using a noninstru- mented assay format that could be used to document follicular development and ovulation outside of the lab- oratory. The purpose of this study was to evaluate the ad- aptation of an instrumented EIA for estrone conjugates (EIC) to a noninstrumented EIA (NEIA). Changes in urinary EIC concentrations measured by NEIA were compared to changes in serum estradiol concentrations in paired daily blood and urine samples throughout ovulatory menstrual cycles. Comparisons were made 2003

Upload: coralie-j

Post on 02-Jan-2017

212 views

Category:

Documents


0 download

TRANSCRIPT

A prototype for ovulation detection: Pros and cons

Bill L. Lasley, PhD, Susan E. Shideler, PhD, and Coralie J. Munro, BSc

Davis, California

A noninstrumented enzyme immunoassay for urinary estrone conjugates was adapted from an instrumented microtiter plate enzyme immunoassay assay. The end point of the assay was a color change from green to clear, which was visible to the unaided eye. The visible color change was adjusted to allow 80 ng l ml estrone conjugates (on the basis of a sample size of 6.5 ,...1 urine) to be distinguished from an infinite dilution without instrumentation. The evaluation of human urine collected from ovulatory ovarian cycles demonstrated that early follicular phase concentrations (35.9 ± 6.8 to 79.4 ± 14.7 nglml, n = 10) produced a dark-green color, whereas late follicular phase concentrations (162.9 ± 20.1 nglml, n = 10) produced no color. Daily urine samples throughout 10 ovulatory ovarian cycles produced parallel profiles when compared to measurements of estradiol in paired blood samples. Complete analysis of the data indicated that ovarian follicular dynamics can be accurately monitored through the noninstrumented analysis of daily estrone conjugates in urine samples. (AM J OSSTET GVNECOL 1991 ;165:2003-7.)

The events associated with ovarian follicular growth and development traditionally have been characterized by the measurement of serum or plasma estradiol. More recently, radioimmunoassay (RIA) measure­ments of urinary metabolites have been shown to be practical replacements for measurements of blood con­centrations of estradiol in human'-3 and nonhuman pri­mates.' Urinary measurements of excreted hormone concentrations have been shown to provide an accurate appraisal of estrogen dynamics during both the normal and the abnormal human menstrual cycle as well as during the periimplantation period in human be­ings3. 5 and macaque monkeys .'

At the present time, efforts are being made to pro­vide nonradiometric assay substitutes for RIA methods. Enzyme immunoassays (EIAs) have been developed for urinary steroid metabolites.6 •8 Urinary analysis of ovar­ian function has advantages over serum analysis be­cause subjects can collect their own samples for lengthy studies and store them by freezing without preserva­tives for analysis at a central site. Enzyme assay formats have advantages over RIAs in that they eliminate toxic and regulated substances, reduce monetary and labor costs of analysis, and enhance the ability to expand study designs to include the monitoring of large num­bers of subjects over relatively large geographic areas. Additionally, enzyme immunoassay formats provide

From the Department of Reproduction, School of Veterinary M edicine, University of California. Supported by The National Institute of International Studies in Nat­ural Family Planning under a cooperative agreement with the U.S . Agency for International Development DPE-3040-AOO-5064-00. Reprint requests: Bill Lasley, PhD, Department of R eproduction, School of Veterinary Medicine, University of California, Davis, CA 95616. 610134954

the basis for the development of noninstrumented as­says that can be used under nonlaboratory conditions. Such assays would allow subjects to collect and analyze their own daily urine samples and, in certain situations, detect the fertile period or anticipate ovulation.

Commercial tests are currently available for the pri­mary metabolites of progesterone, but there are few, if any, tests for impending ovulation. Most tests, like those signaling the luteinizing hormone (LH) surge, or change in the estrogen / progesterone ratio, allow de­tection or confirmation of ovulation only after the fer­tile period has begun. Such indicators can be used for fertility enhancement, but they have little value as an aid to prospectively avoid conception.

The best indicator of the fertile period is the initiation of the preovulatory estrogen rise. This event coincides with the selection of the ovarian follicle , which will ovu­late 3 to 6 days later. If the initial rise in serum estradiol is predictive of ovulation by 5 days, then it would follow that urinary metabolites also can be used for the same prediction-assuming a minimal lag time between changes in serum estradiol and urinary metabolites. We have demonstrated a preovulatory rise in both serum and urinary estrogen,s but have not correlated the events in urine to those in serum using a noninstru­mented assay format that could be used to document follicular development and ovulation outside of the lab­oratory.

The purpose of this study was to evaluate the ad­aptation of an instrumented EIA for estrone conjugates (EIC) to a noninstrumented EIA (NEIA). Changes in urinary EIC concentrations measured by NEIA were compared to changes in serum estradiol concentrations in paired daily blood and urine samples throughout ovulatory menstrual cycles. Comparisons were made

A prototype for ovulation detection: Pros and cons

Bill L. Lasley, PhD, Susan E. Shideler, PhD, and Coralie J. Munro, BSe

Davis, California

A noninstrumented enzyme immunoassay for urinary estrone conjugates was adapted from an instrumented microtiter plate enzyme immunoassay assay. The end point of the assay was a color change from green to clear, which was visible to the unaided eye. The visible color change was adjusted to allow 80 nglml estrone conjugates (on the basis of a sample size of 6.5 fJol urine) to be distinguished from an infinite dilution without instrumentation. The evaluation of human urine collected from ovulatory ovarian cycles demonstrated that early follicular phase concentrations (35.9 ± 6.8 to 79.4 ± 14.7 nglml, n = 10) produced a dark-green color, whereas late follicular phase concentrations (162.9 ± 20.1 nglml, n = 10) produced no color. Daily urine samples throughout 10 ovulatory ovarian cycles produced parallel profiles when compared to measurements of estradiol in paired blood samples. Complete analysis of the data indicated that ovarian follicular dynamics can be accurately monitored through the noninstrumented analysis of daily estrone conjugates in urine samples. (AM J OBSTET GVNECOL 1991;165:2003-7.)

The events associated with ovarian follicular growth and development traditionally have been characterized by the measurement of serum or plasma estradiol. More recently, radioimmunoassay (RIA) measure­ments of urinary metabolites have been shown to be practical replacements for measurements of blood con­centrations of estradiol in human 1·3 and nonhuman pri­mates.' Urinary measurements of excreted hormone concentrations have been shown to provide an accurate appraisal of estrogen dynamics during both the normal and the abnormal human menstrual cycle as well as during the periimplantation period in human be­ings3.5 and macaque monkeys.'

At the present time, efforts are being made to pro­vide nonradiometric assay substitutes for RIA methods. Enzyme immunoassays (EIAs) have been developed for urinary steroid metabolites.6 •s Urinary analysis of ovar­ian function has advantages over serum analysis be­cause subjects can collect their own samples for lengthy studies and store them by freezing without preserva­tives for analysis at a central site. Enzyme assay formats have advantages over RIAs in that they eliminate toxic and regulated substances, reduce monetary and labor costs of analysis, and enhance the ability to expand study designs to include the monitoring of large num­bers of subjects over relatively large geographic areas. Additionally, enzyme immunoassay formats provide

From the Department of Reproduction, School ofVete1-inary Medicine, University of California. Supported by The National Institute of International Studies in Nat­ural Family Planning under a cooperative agreement with the U.S. Agency for International Development DPE-3040-AOO-5064-00. Reprint requests: Bill Lasley, PhD, Department of Reproduction, School of Veterinary Medicine, University of California, Davis, CA 95616. 610134954

the basis for the development of noninstrumented as­says that can be used under nonlaboratory conditions. Such assays would allow subjects to collect and analyze their own daily urine samples and, in certain situations, detect the fertile period or anticipate ovulation.

Commercial tests are currently available for the pri­mary metabolites of progesterone, but there are few, if any, tests for impending ovulation. Most tests, like those signaling the luteinizing hormone (LH) surge, or change in the estrogen/progesterone ratio, allow de­tection or confirmation of ovulation only after the fer­tile period has begun. Such indicators can be used for fertility enhancement, but they have little value as an aid to prospectively avoid conception.

The best indicator of the fertile period is the initiation of the preovulatory estrogen rise. This event coincides with the selection of the ovarian follicle, which will ovu­late 3 to 6 days later. If the initial rise in serum estradiol is predictive of ovulation by 5 days, then it would follow that urinary metabolites also can be used for the same prediction-assuming a minimal lag time between changes in serum estradiol and urinary metabolites. We have demonstrated a preovulatory rise in both serum and urinary estrogen,s but have not correlated the events in urine to those in serum using a noninstru­mented assay format that could be used to document follicular development and ovulation outside of the lab­oratory.

The purpose of this study was to evaluate the ad­aptation of an instrumented EIA for estrone conjugates (EIC) to a noninstrumented EIA (NEIA). Changes in urinary EIC concentrations measured by NEIA were compared to changes in serum estradiol concentrations in paired daily blood and urine samples throughout ovulatory menstrual cycles. Comparisons were made

2003

2004 Lasley et al.

between estrogen concentrations In daily serum as­sessed by RIA and urine samples assessed by both in­strumented and noninstrumented EIAs to determine the timing before ovulation of the first estrogen rise in each.

Material and methods

Subjects, samples, and previous analysis. The se­rum and urine samples used in this study were a subset of those described in a previous report." Subjects and collection of samples are described in detail in the pre­vious report. Briefly, women between the ages of 23 and 40 years with normal menstrual cycles were re­cruited for the study. Daily early morning urine sam­ples and midmorning blood samples were collected throughout one complete menstrual cycle . Samples were frozen and stored until evaluation. After evalu­ation, remaining urine samples were refrozen and stored for approximately 1 year. Urine samples for complete menstrual cycles of six subjects were then reanalyzed by the NEIA reported here . These ovarian cycles were previously shown to be ovulatory by the assessment of serum LH, follicle-stimulating hormone (FSH), estradiol (E2 ), and progesterone (Po) using com­mercial kits purchased from Diagnostic Products Inc. (Los Angeles, Calif.).8 In addition, cycles were analyzed for urinary EIC and progesterone (PdG) by validated laboratory methods described by Munro et al.8 Creat­inine was measured by the method of Taussky (1969) to adjust for the differences in urine concentration. The day of the midcycle LH peak was used as a ref­erence for the day of ovulation and a sustained rise of serum progesterone for 12 to 15 days was observed in all menstrual cycles studied.

Estrone conjugate noninstrumented enzyme im­munoassay. The EIC NEIA is similar in format to the EIC EIA described by Czekala et al. 6 and uses most of the materials reported by Munro et al ." Three major changes were introduced to the microtiter plate EIC EIA format. High-binding star tubes (Nunc Maxisorp star tubes, Nunc Inc., Naperville, III.) were substituted for the microtiter plate as a solid matrix for the assay to maximize the color change during the midfollicular phase when the estrogen production of the dominant follicle signaled the beginning of the fertile period. Sec­ond, the estrone conjugate enzyme label was altered to

maximize the dose response dynamic by eliminating the glucuronide moiety. This steepened the slope and pro­vided an "all-or-none" or "binary" signal to be gener­ated over a small change in concentration of standard or unknown. The 50% binding point of the standard curve of the EIC NEIA with estrone horseradish per­oxidase as the enzyme competitor was 400 pgltube (or

December 1991 Am J Obstet Gynecol

80 ng/ ml on the basis of a sample size of 6.5 flol urine) of estrone-3-glucuronide. Third, small Whatman No. 1 filter paper pads (7.5 mm diameter) were used to measure and transfer urine samples. Pads were dried completely and stored at 4° C until assayed. Several pad samples of individual urines were analyzed to test the reliability of sample volume measurement as well as the shelf life of the sample dried onto the absorbent pad. Reliability was comparable to micropipettors (6.5 ± 0.3 flol, n = 12, coefficient of variation 1.6%); samples could be stored up to several weeks with refrigeration before analysis without loss in reactivity: one cycle was ana­lyzed by EIC NEIA on four different occasions by use of sample pads that had been stored for more than 6 months at room temperature, with no change in results.

The EIC NEIA was performed by coating star tubes with 0.5 ml of antibody R522.8 After a minimum in­cubation of 18 to 24 hours, the tubes were rinsed with EIA wash. A clean filter paper pad was soaked in a urine sample for 1 to 3 minutes, blotted dry, and placed into 0.5 ml of the conjugate solution added to the pre­coated star tube. After a 60-minute incubation the con­tents of the tube was discarded , the tube was rinsed with EIA wash, and 0.5 ml of substrate was added. The visual color change was evaluated quantitatively by measuring the optical density of a 0 .1 ml aliquot trans­ferred to a micro titer plate and read on a Dynatech MR600 spectrophotometer (405 nm) .

Statistics. Composite cycles were derived from mean daily hormone values of the individual subjects and were aligned according to the serum preovulatory LH / FSH peak. CUSUM analysis, as described by Royston9 was used to describe the pattern of urinary estrogen concentrations before ovulation. The "first rise" of serum E2 or urinary EIC was defined as that day before the LH surge on which the mean for all cycles was greater than a reference value. The refer­ence value was defined as "baseline plus 0.3." The "baseline," in turn, was defined as the mean of the first 6 days of each individual cycle expressed as a natural log. A significant rise was defined as that day before the LH / FSH peak on which the value exceeded the "de­cision interval" (0.6).

Results

When compared with serum E2 measurements, the urinary EIC EIA values lagged behind by 0 .7 day, in­dicating that less than 1 day was lost when the urinary dynamic of the EIC was measured in place of serum E2. The composite urinary EIC profile determined by EIA and the serum E2 (concentration measured in pi­comoles per liter) profile determined by RIA for the same 10 cycles are shown in Fig. 1. Although composite

2004 Lasley et al.

between estrogen concentrations m daily serum as­sessed by RIA and urine samples assessed by both in­strumented and noninstrumented EIAs to determine the timing before ovulation of the first estrogen rise in each.

Material and methods

Subjects, samples, and previous analysis. The se­rum and urine samples used in this study were a subset of those described in a previous report." Subjects and collection of samples are described in detail in the pre­vious report. Briefly, women between the ages of 23 and 40 years with normal menstrual cycles were re­cruited for the study. Daily early morning urine sam­

ples and midmorning blood samples were collected throughout one complete menstrual cycle. Samples were frozen and stored until evaluation. After evalu­ation, remaining urine samples were refrozen and

stored for approximately 1 year. Urine samples for complete menstrual cycles of six subjects were then reanalyzed by the NEIA reported here. These ovarian cycles were previously shown to be ovulatory by the assessment of serum LH, follicle-stimulating hormone (FSH), estradiol (E2)' and progesterone (Po) using com­mercial kits purchased from Diagnostic Products Inc. (Los Angeles, Calif.)." In addition, cycles were analyzed for urinary EIC and progesterone (PdG) by validated laboratory methods described by Munro et al." Creat­inine was measured by the method of Taussky (1969) to adjust for the differences in urine concentration. The day of the midcycle LH peak was used as a ref­erence for the day of ovulation and a sustained rise of serum progesterone for 12 to 15 days was observed in all menstrual cycles studied.

Estrone conjugate noninstrumented enzyme im­munoassay. The EIC NEIA is similar in format to the EIC EIA described by Czekala et al." and uses most of the materials reported by Munro et aU Three major changes were introduced to the microtiter plate EIC EIA format. High-binding star tubes (Nunc Maxisorp star tubes, Nunc Inc., Naperville, Ill.) were substituted for the micro titer plate as a solid matrix for the assay to maximize the color change during the mid follicular phase when the estrogen production of the dominant follicle signaled the beginning of the fertile period. Sec­ond, the estrone conjugate enzyme label was altered to maximize the dose response dynamic by eliminating the glucuronide moiety. This steepened the slope and pro­vided an "all-or-none" or "binary" signal to be gener­ated over a small change in concentration of standard or unknown. The 50% binding point of the standard curve of the EIC NEIA with estrone horseradish per­oxidase as the enzyme competitor was 400 pgltube (or

December 1991 Am J Obstet Gynecol

80 ng/ml on the basis of a sample size of 6.5 J.d urine) of estrone-3-glucuronide. Third, small Whatman No. 1 filter paper pads (7.5 mm diameter) were used to

measure and transfer urine samples. Pads were dried completely and stored at 4 0 C until assayed. Several pad samples of individual urines were analyzed to test the reliability of sample volume measurement as well as the shelf life of the sample dried onto the absorbent pad. Reliability was comparable to micropipettors (6.5 ± 0.3 J.LI, n = 12, coefficient of variation 1.6%); samples could be stored up to several weeks with refrigeration before analysis without loss in reactivity: one cycle was ana­lyzed by EIC NEIA on four different occasions by use of sample pads that had been stored for more than 6 months at room temperature, with no change in results.

The EIC NEIA was performed by coating star tubes with 0.5 ml of antibody R522.H After a minimum in­

cubation of 18 to 24 hours, the tubes were rinsed with EIA wash. A clean filter paper pad was soaked in a urine sample for 1 to 3 minutes, blotted dry, and placed into 0.5 ml of the conjugate solution added to the pre­coated star tube. After a 60-minute incubation the con­tents of the tube was discarded, the tube was rinsed with EIA wash, and 0.5 ml of substrate was added. The visual color change was evaluated quantitatively by measuring the optical density of a 0.1 ml aliquot trans­ferred to a micro titer plate and read on a Dynatech MR600 spectrophotometer (405 nm).

Statistics. Composite cycles were derived from mean daily hormone values of the individual subjects and were aligned according to the serum preovulatory LH/FSH peak. CUSUM analysis, as described by Royston9 was used to describe the pattern of urinary estrogen concentrations before ovulation. The "first rise" of serum E2 or urinary EIC was defined as that day before the LH surge on which the mean for all cycles was greater than a reference value. The refer­ence value was defined as "baseline plus 0.3." The "baseline," in turn, was defined as the mean of the first 6 days of each individual cycle expressed as a natural log. A significant rise was defined as that day before the LH/FSH peak on which the value exceeded the "de­cision interval" (0.6).

Results

When compared with serum E2 measurements, the urinary EIC EIA values lagged behind by 0.7 day, in­dicating that less than 1 day was lost when the urinary dynamic of the EIC was measured in place of serum E2 • The composite urinary EIC profile determined by EIA and the serum E2 (concentration measured in pi­comoles per liter) profile determined by RIA for the same 10 cycles are shown in Fig. 1. Although composite

Volume 165 Number 6, Part 2

50

l&J z Z

~ 40 a: 0 0 E

~ 30 E C

~ 20 :::» 1 8 l&J z 10

~ 0 -20 -15 -10

Prototype for ovulation detection 2005

1200

900

S 0 E Q.

600 ~

I 300

0 -5 0 5 10 15

DAYS PRE AND POST LH PEAK

Fig. 1. Daily mean serum estradiol and urinary EIC concentrations in 10 ovulatory menstrual cycles. (From Munro CJ , Stabenfeldt GH , CragunJR, Addiego LA, OverstreetJW, Lasley BL. Clin Chern 1991 ;37 :838-44, with permission from Clinical Chemistry. )

peak values for both EIC and E2 occurred on day 0, four of 10 cycles had E2 peaks on day - I with an EIC peak on day 0, and two had E2 peaks on day 0 and EIC peaks on day + I. As previously reported, serum E2 values showed a postovulatory nadir on day + 2, and EIC values in urine showed a postovulatory nadir on day + 4." Both serum E" and urine EIC concentrations increased during the luteal phase , remaining at about 50% of peak values until approximately day + 10 be­fore declining to early follicular phase levels. The mean urinary EIC profile was parallel to the mean serum E" profile (r = 0,88, P < 0.001).

Estrone conjugate evaluations of human urine by EIA indicated that EIC concentrations in the early fol­licular phase ranged from 35.9 ± 6.8 to 70.4 ± 14.7

ng/ml , whereas late follicular phase concentrations were as high as 162 ± 20, I ng / ml. For the EIC NEIA to be of diagnostic use, a significant color change had to occur at the periovulatory estrogen excursion (i ,e., at 70.4 ± 14.7 ng / ml urine). The high-binding star tubes showed a distinct visual color change at approx­imately 70 to 80 ng / ml based on an absorbent pad sample of 6.5 I.d urine, Under these conditions, early follicular phase concentrations exhibited dark-green

coloration, whereas late follicular phase and ovulatory concentrations were indicated by no color. In general, the dark-green color progressively lightened over two or three samples 3 to 4 days before the LH/FSH peak and was nearly colorless in all cycles on the day of the EIC peak, as determined by instrumented EIA.

When six ovarian cycles evaluated by NEIA were assessed quantitatively by transferral of both the stan­dard curve and unknowns from the star tubes to a microtiter plate for reading the results on a spectro­photometer, a composite profile was obtained that was identical to that produced by the original EIC EIA for IO cycles. The individual profiles of serum E" and uri­nary EIC as measured by NEIA-like those measured by instrumented EIC EIA-were parallel (r = 0.88,

P < 0.01). Statistical analysis by CUSUM for the six menstrual

cycles evaluated for both E2 by RIA and EIC by EIA and EIC NEIA are presented in Table I. The "first rise" is defined in CUSUM analysis as the day before the LH surge that the CUSUM becomes positive-i.e. , above the reference (baseline + 0.3). Baseline is calculated as the mean of the first 6 days of each cycle and was expressed as log", and significance is determined as the

Volume 165 Number 6, Part 2

50

1.&.1 z z 8 40 It: 0 "0 E

~ 30 E c

~ 20 :::)

~ 8 1.&.1 z 10

~ 0 -20 -15 -10

Prototype for ovulation detection 2005

1200

900

S 0 E a.

600 ~

I 300

0 -5 0 5 10 15

DAVS PRE AND POST LH PEAK

Fig. 1. Daily mean serum estradiol and urinary EIC concentrations in 10 ovulatory menstrual cycles. (From Munro Cj, Stabenfeldt GH, CragunJR, Addiego LA, OverstreetJW, Lasley BL. Clin Chern 1991 ;37:838-44, with permission from Clinical Chemistry.)

peak values for both ElC and E2 occurred on day 0, four of 10 cycles had E2 peaks on day - 1 with an ElC peak on day 0, and two had E2 peaks on day 0 and EIC peaks on day + 1. As previously reported, serum E2 values showed a postovulatory nadir on day + 2, and ElC values in urine showed a postovulatory nadir on day +4." Both serum E2 and urine EIC concentrations increased during the luteal phase, remaining at about 50% of peak values until approximately day + 10 be­fore declining to early follicular phase levels. The mean urinary EIC profile was parallel to the mean serum E2 profile (r = 0.88, P < 0.001).

Estrone conjugate evaluations of human urine by EIA indicated that EIC concentrations in the early fol­licular phase ranged from 35.9 ± 6.8 to 70.4 ± 14.7 ng/ml, whereas late follicular phase concentrations were as high as 162 ± 20.1 ng/ml. For the EIC NEIA to be of diagnostic use, a significant color change had to occur at the periovulatory estrogen excursion (i.e., at 70.4 ± 14.7 ng/ml urine). The high-binding star tubes showed a distinct visual color change at approx­imately 70 to 80 ng/ml based on an absorbent pad sample of 6.5 fl.1 urine. Under these conditions, early follicular phase concentrations exhibited dark-green

coloration, whereas late follicular phase and ovulatory concentrations were indicated by no color. In general, the dark-green color progressively lightened over two or three samples 3 to 4 days before the LH/FSH peak and was nearly colorless in all cycles on the day of the EIC peak, as determined by instrumented EIA.

When six ovarian cycles evaluated by NEIA were assessed quantitatively by transferral of both the stan­dard curve and unknowns from the star tubes to a microtiter plate for reading the results on a spectro­photometer, a composite profile was obtained that was identical to that produced by the original ElC EIA for 10 cycles. The individual profiles of serum E2 and uri­nary ElC as measured by NEIA-like those measured by instrumented ElC EIA-were parallel (r = 0.88, P < 0.01).

Statistical analysis by CUSUM for the six menstrual cycles evaluated for both E2 by RIA and ElC by EIA and ElC NEIA are presented in Table I. The "first rise" is defined in CUSUM analysis as the day before the LH surge that the CUSUM becomes positive-i.e., above the reference (baseline + 0.3). Baseline is calculated as the mean of the first 6 days of each cycle and was expressed as loge, and significance is determined as the

2006 Lasley et al. December 199 1 Am J Obstet Gynecol

Table I. CUSUM results for serum E2(pg / ml), urinary ElC as measured by instrumented enzyme immunoassay and indexed by creatinine (EIA/ CR), and urinary E 1 C as measured by the noninstrumented format and indexed by creatinine (NEIA/CR)

Subject No.

1 2 3 4 5 6 Range M ean

Serum E2 First rise -3 -8 -8 -7 - s -s -3 to - 8 -6 Significance -2 -6 - 6 -6 -4 - 3 -2 to - 6 -4.S

Urine EIC EIA /CR First rise -2 -8 - 6 -7 -4 - 6 -2 to - 7 -s.s Significance -I -6 -S -6 - 2 - 3 -I to - 6 -3.8

Urine EIC NEIA First rise 0 -7 -6 -6 02 - 4 o to - 7 -5 Significance +2 - I -6 -6 - 1 -3 +2 to - 6 -3.4

(From Munro CJ, Stabenfeldt GH, Cragun JR, Addiego LA, Overstreet JW, Lasley BL. Clin Chern 1991;37:838-44, with permission from Clinical Chemistry.)

day before the LH surge that the CUSUM exceeds the decision interval (0.6). With these CUSUM criteria, daily serum E2 concentrations accurately predicted the fertile period in five (86%) of the six cycles analyzed. Urinary ElC measured by microtiter plate EIA and in­dexed by creatinine gave similar results when the 1 day lag in excretion compared to serum E2 secretion was considered, with a mean detection of 5.5 days , confirm­ing that urinary ElC can be used in place of serum E2 to predict the fertile period. Essentially the same results were obtained with the ElC NEIA as with the micro titer plate ElC EIA without indexing by creatinine.

Comment

The present study was conducted on the basis of historic data that indicate that the increase of serum E2

concentrations accurately and consistently predict the occurrence of impending ovulation. The data pre­sented here support that indication and demonstrate that, with early morning samples , changes in urinary estrogen excretion during the normal ovarian cycle can also be used to predict ovulation. In addition, urinary estrogen metabolites can be detected with an NEIA and that these measures also predict the ovulatory event. Although the numbers of observations are limited, a consistent rise of urinary ElC was detected as early as 6 days and as late as 2 days before the midcycle LH peak, which was comparable to the rises detected when E2 was measured in paired blood samples . In general, E2 measurements provided an earlier and more con­sistent prediction of the LH surge, but the advantages of the ElC NEIA and the use of urine more than com­pensate for the small difference lost in predictiveness. The immediate use of this assay is to provide a self­

evaluation of the major ovarian follicular events and predict ovulation several days in advance.

In five of the six cycles monitored by ElC NEIA, the preovulatory estrogen rise in urine was detected at ap­proximately the time of the beginning of the fertile period, which is considered to be 5 days before ovu­lation or 4 days before the LH surge. ' It would appear therefore that this technique has the potential for pre­diction of ovulation by several days. This information could be used for either the timing of propitious in­semination or for the anticipation of the fertile period so as to reduce the risk of pregnancy at this time.

When ElC values were combined with the measure­ment of urinary progesterone and luteinizing hormone metabolite measurements, the detection of the ElC rise during the follicular phase provided the ability to com­pletely and comprehensively monitor all major ovarian events of the complete menstrual cycle. The ability to generate such information may in turn provide the basis for more detailed future studies of ovarian func­tion in contexts in which biologic samples other than urine cannot be collected and stored for later evalua­tion. The ElC NEIA presented here is particularly use­ful when overall cycle lengths are variable or when even a general estimate of the day of ovulation cannot be made.

The inherent variations in urine concentration, which are compensated for in the laboratory by index­ing each hormone value by the creatinine concentra­tion, can, in the extreme case, contribute to false in­formation. Whereas this is not a problem with the measurement of urinary luteinizing hormone or pro­gesterone metabolites in which large increases of signal is exhibited over the baseline values, the relatively small baseline to peak ratio of the ElC profile is perturbed by variations in urine concentration that can be seen under normal conditions : false-positive and false-neg­ative ElC NEIA results are to be expected when urine

2006 Lasley et al. December 1991 Am J Obslet Gynecol

Table I. CUSUM results for serum E~(pg/ml), urinary ElC as measured by instrumented enzyme immunoassay and indexed by creatinine (EIA/CR), and urinary ElC as measured by the noninstrumented format and indexed by creatinine (NEIA/CR)

Subject No.

1 2 3 4 5 6 Range Mean

Serum E, First rise -3 -8 -8 -7 -s -s -3 to -8 -6 Significance -2 -6 -6 -6 -4 -3 -2to -6 -4.5

Urine EIC EIA/CR First rise -2 -8 -6 -7 -4 -6 -2 to -7 -S.S Significance -I -6 -s -6 -2 -3 -1 to -6 -3.8

Urine EIC NEIA First rise 0 -7 -6 -6 02 -4 OtD -7 -5 Significance +2 -I -6 -6 -I -3 +2 to -6 -3.4

(From Munro Cj, Stabenfeldt GH, Cragun JR, Addiego LA, Overstreet JW, Lasley BL. Clin Chern 1991;37:838-44, with permission from Clinical Chemistry.)

day before the LH surge that the CUSUM exceeds the decision interval (0.6). With these CUSUM criteria, daily serum E, concentrations accurately predicted the fertile period in five (86%) of the six cycles analyzed. Urinary ElC measured by microtiter plate EIA and in­dexed by creatinine gave similar results when the 1 day lag in excretion compared to serum E" secretion was considered, with a mean detection of 5.5 days, confirm­ing that urinary ElC can be used in place of serum E, to predict the fertile period. Essentially the same results were obtained with the ElC NEIA as with the micro titer plate ElC EIA without indexing by creatinine.

Comment

The present study was conducted on the basis of historic data that indicate that the increase of serum E, concentrations accurately and consistently predict the occurrence of impending ovulation. The data pre­sented here support that indication and demonstrate that, with early morning samples, changes in urinary estrogen excretion during the normal ovarian cycle can also be used to predict ovulation. In addition, urinary estrogen metabolites can be detected with an NEIA and that these measures also predict the ovulatory event. Although the numbers of observations are limited, a consistent rise of urinary ElC was detected as early as 6 days and as late as 2 days before the midcycle LH peak, which was comparable to the rises detected when E, was measured in paired blood samples. In general, E2 measurements provided an earlier and more con­sistent prediction of the LH surge, but the advantages of the ElC NEIA and the use of urine more than com­pensate for the small difference lost in predictiveness. The immediate use of this assay is to provide a self­

evaluation of the major ovarian follicular events and predict ovulation several days in advance.

In five of the six cycles monitored by ElC NEIA, the preovulatory estrogen rise in urine was detected at ap­proximately the time of the beginning of the fertile period, which is considered to be 5 days before ovu­lation or 4 days before the LH surge.' It would appear therefore that this technique has the potential for pre­diction of ovulation by several days. This information could be used for either the timing of propitious in­semination or for the anticipation of the fertile period so as to reduce the risk of pregnancy at this time.

When ElC values were combined with the measure­ment of urinary progesterone and luteinizing hormone metabolite measurements, the detection of the ElC rise during the follicular phase provided the ability to com­pletely and comprehensively monitor all major ovarian events of the complete menstrual cycle. The ability to generate such information may in turn provide the basis for more detailed future studies of ovarian func­tion in contexts in which biologic samples other than urine cannot be collected and stored for later evalua­tion. The ElC NEIA presented here is particularly use­ful when overall cycle lengths are variable or when even a general estimate of the day of ovulation cannot be made.

The inherent variations in urine concentration, which are compensated for in the laboratory by index­ing each hormone value by the creatinine concentra­tion, can, in the extreme case, contribute to false in­formation. Whereas this is not a problem with the measurement of urinary luteinizing hormone or pro­gesterone metabolites in which large increases of signal is exhibited over the baseline values, the relatively small baseline to peak ratio of the ElC profile is perturbed by variations in urine concentration that can be seen

under normal conditions: false-positive and false-neg­ative ElC NEIA results are to be expected when urine

Volume 165 Number 6, Part 2

osmolality is either high or low, respectively. A safe­guard against such errors would be to regulate fluid intake and therefore normalize urine production dur­ing testing periods. Alternatively, urines could be tested with a simple device simultaneous to the NEIA to en­sure that each urine concentration falls within a pre­scribed range. It seems likely that this aspect of estrogen monitoring will require the development of a nonin­strumented test for urine osmolality in the near future.

REFERENCES I. Wright K, Collins DC, Musey PI, Preedy JKR. Direct ra­

dioimmunoassay of specific urinary glucuronates in normal men and non-pregnant women. Steroids 1978;31 :407-26.

2. Branch CM, Collins PO, Kilpatrick MJ, Manning PA, Pike JM, Tyler JPP. The concentration of urinary estrone-3-glucuronide, LH, and pregnanediol-3-alpha-glucuronide as indices of ovarian function . Acta Endocrinol (Copenh) 1979;90:336-48.

3. Lasley BL, Stabenfeldt GH, Overstreet JW, Hanson FW, Czekala NM, Munro CJ. Urinary hormone levels at the time of ovulation and implantation. Ferti! Steril 1985;43:681-7.

Prototype for ovulation detection 2007

4. Monfort SL, Hess DL, Shideler SE, Samuels SJ, Hendrickx AG, Lasley BL. Comparison of serum estradiol to urinary estrone conjugates in the rhesus macaque (Macaca mulatta). Bioi Reprod 1987;37:832-7.

5. Shideler SE, DeVane GW, Kalra PS, Benirschke K, Lasley BL. Ovarian-pituitary hormone interactions during the perimenopause. Maturitas 1989; II :331-9.

6. Czekala NM, Galluser S, Meier J, Lasley BL. The devel­opment and application of and enzyme assay for estrone conjugates. Zoo Bioi 1986;6:1-6.

7. Shideler SE, Tell LA, Owiti GE, Laughlin L, Chatterton R, Lasley BL. The relationship of serum estradiol and pro­gesterone concentrations to the enzyme immunoassay mea­surements of urinary estrone conjugates and immuno­reactive pregnanediol-3-glucuronide in Macaca mulatta. AmJ PrimatoI1990;22:113-22.

8. Munro Cj, Stabenfeldt GH, Cragun JR, Addiego LA, Overstreet JW, Lasley BL. Relationship of serum estradiol and progesterone concentrations to the excretion profiles of their major urinary metabolites as measured by enzyme immunoassay and radioimmunoassay. Clin Chern 1991;37:838-44.

9. Royston JP. Statistical approaches to the prediction and detection of ovulation: detecting the signal among the noise. In: Jeffcoate SL, ed. Ovulation: methods for its pre­diction and detection. New York: John Wiley & Sons, 1983.

Volume 165 Number 6, Part 2

osmolality is either high or low, respectively. A safe­guard against such errors would be to regulate fluid intake and therefore normalize urine production dur­ing testing periods. Alternatively, urines could be tested with a simple device simultaneous to the NEIA to en­sure that each urine concentration falls within a pre­scribed range. It seems likely that this aspect of estrogen monitoring will require the development of a nonin­strumented test for urine osmolality in the near future.

REFERENCES 1. Wright K, Collins DC, Musey PI, Preedy JKR. Direct ra­

dioimmunoassay of specific urinary glucuronates in normal men and non-pregnant women. Steroids 1978;31:407-26.

2. Branch CM, Collins PO, Kilpatrick MJ, Manning PA, Pike JM, Tyler JPP. The concentration of urinary estrone-3-glucuronide, LH, and pregnanediol-3-alpha-glucuronide as indices of ovarian function. Acta Endocrinol (Copen h) 1979;90:336-48.

3. Lasley BL, Stabenfeldt GH, Overstreet JW, Hanson FW, Czekala NM, Munro CJ. Urinary hormone levels at the time of ovulation and implantation. Ferti! Steril 1985;43:681-7.

Prototype for ovulation detection 2007

4. Monfort SL, Hess DL, Shideler SE, Samuels SJ, Hendrickx AG, Lasley BL. Comparison of serum estradiol to urinary estrone conjugates in the rhesus macaque (Macaca mulatta). Bioi Reprod 1987;37:832-7.

5. Shideler SE, DeVane GW, Kalra PS, Benirschke K, Lasley BL. Ovarian-pituitary hormone interactions during the perimenopause. Maturitas 1989; 11 :331-9.

6. Czekala NM, Galluser S, Meier J, Lasley BL. The devel­opment and application of and enzyme assay for estrone conjugates. Zoo Bioi 1986;6:1-6.

7. Shideler SE, Tell LA, Owiti GE, Laughlin L, Chatterton R, Lasley BL. The relationship of serum estradiol and pro­gesterone concentrations to the enzyme immunoassay mea­surements of urinary estrone conjugates and immuno­reactive pregnanediol-3-glucuronide in Macaca mulatta. Am J Primatol 1990;22: 113-22.

8. Munro Cj, Stabenfeldt GH, Cragun JR, Addiego LA, Overstreet JW, Lasley BL. Relationship of serum estradiol and progesterone concentrations to the excretion profiles of their major urinary metabolites as measured by enzyme immunoassay and radioimmunoassay. Clin Chern 1991;37:838-44.

9. Royston JP. Statistical approaches to the prediction and detection of ovulation: detecting the signal among the noise. In: Jeffcoate SL, ed. Ovulation: methods for its pre­diction and detection. New York: John Wiley & Sons, 1983.