Journal of Assisted Reproduction and Genetics, Vol. 11, No. 8, 1994 CLINICAL ASSISTED REPRODUCTION

Relationship Between Urinary Estrone Conjugates as Measured by Enzyme Immunoassay and Serum Estradiol in Women Receiving Gonadotropins for in Vitro Fertilization

MICHAEL M. ALPER, 1'3 LISA HALVORSON, 1 BILL LASLEY, 2 and JOSEPH MORTOLA 1

Submitted: October 14, 1994 Accepted: December 1, 1994

Objective: Our purpose was to determine whether urinary estrone conjugates (E1C) as measured by enzyme immu- noassay correlate with serum estradiol (E2) in women undergoing controlled ovarian hyperstimulation with hu- man menopausal gonadotropins. Design: This was a prospective, clinical study. Setting: The study took place in an outpatient, university- affiliated in vitro fertilization (IVF) unit. Interventions: First morning urine samples were analyzed for EIC using a competitive solid-phase microtiter en- zyme immunoassay and the value was corrected for uri- nary creatinine (E1C/Cr). The value was compared to morning serum E2 as determined by radioimmunoassay. Results: Mean E2 and EIC/Cr levels demonstrated a sim- ilar pattern on the days before hCG administration. The correlation between EIC and E2 was 0.85 (P < 0.0001). Furthermore, the correlation between the number o f fol- licles greater than 12 mm was as high for E1C/Cr (p = 0.71, P < 0.001) as it was fo rE2 (p = 0.74, P < 0.0001). Conclusions: Urinary EIC/Cr levels in women receiving hMG correlate with serum E2. Further studies are nec- essary to determine whether E1C is clinically useful to predict ovarian hyperstimulation syndrome.

KEY WORDS: estrogens; in vitro fertilization; gonadotropins; estrone.

1 Beth Israel Hospital and Harvard Medical School, Boston, Massachuse t t s .

2 Depar tment of Obstetr ics and Gynecology, School of Medi- cine, University of California at Davis, Davis, California.

3 To w h o m correspondence should be addressed at One Brook- line Place, Suite 602, Brookline, Massachuse t t s 02146.

INTRODUCTION

Ovarian hyperstimulation syndrome is a potentially serious complication of controlled ovarian hyper- stimulation (COH) using human menopausal gonad- otropins (hMG). In order to minimize the risk of this syndrome estrogen determinations and ovarian ul- trasound measurements are routinely employed. Serum E2 has been repeatedly shown to have pre- dictive value to determine the likelihood developing ovarian hyperstimulation syndrome (1). However, repeated venopuncture required to obtain serial E2 levels represents a definite inconvenience and stress for the patient. Therefore a urinary assay that would predict the risk of ovarian hyperstimulation syndrome would be desirable.

The primary metabolites of E2 are estrone-3- glucuronide and estrone sulfate, which are excreted in the urine. Previous reports have demonstrated that the urinary metabolites of E2 are positively cor- related with serum E2 concentrations in untreated, ovulating women (2,3). Munro and co-workers more recently reported on the development of an enzyme immunoassay (EIA) to measure urinary es- trone conjugates (E1C) in normally ovulating women and detected an excellent correlation with E1C by RIA as well as serum E2 (4). The purpose of this report was to determine if E1C measured by EIA correlated with serum E2 in women undergo- ing ovarian hyperstimulation with gonadotropin therapy.

405 1058-0468/94/0900-0405507.00/0 �9 1994 Plenum Publishing Corporation

406 ALPER, HALVORSON, LASLEY, AND MORTOLA

MATERIALS AND METHODS

Subjects and Sample Collection

Thirty-three healthy women undergoing IVF were recruited to participate in the study. Of these, 25 patients completed the urine collection ade- quately and had three or more urine samples (range, three to six) available for analysis. These 25 pa- tients, ages 29 to 39 years, were used in the analy- sis. Appropriate informed consent was obtained.

The protocol for COH during the IVF cycle em- ployed leuprolide acetate (Lupron; Tap Pharmaceu- ticals Inc., Deerfield, IL) beginning on day 1 of the menstrual cycle. Human menopausal gonadotro- pins (hMG; Pergonal; Serono Laboratories, Nor- well, MA), 150-300 IU, were administered begin- ning on day 2 of the cycle for 5 days. The dose was then adjusted based upon periodic follicle size as determined by vaginal ultrasound and serum E2 measurements. Blood for E2 determination was drawn between 0700 and 1130 beginning on day 6 of the cycle, and the serum was separated and ana- lyzed later that day. At that time subjects were asked to collect their first morning urine sample in three separate test tubes and place the samples im- mediately in their freezer (at about - 20~ without preservatives. The urinary samples were kept fro- zen until transported to the laboratory at the end of the study.

Immunoassay

SERUM E2 ASSAY. Serum E2 was measured by a commercially available solid-phase 125I radioim- munoassay (Coat-A-Count; Diagnostic Products Corporation, Los Angeles, CA) according to the manufacturer's protocol. The intraassay and in- terassay coefficients of variation (CV) were 4.0 and 4.2%, respectively.

URINARY E1C. E1C was measured by a com- petitive, microtiter, solid-phase enzyme immunoas- say (EIA) previously described in detail (4). The assay measures estrone-3-glucuronide (100% cross- react ion), es t rone-3-sulfate (66.5%), estrone (236%), E2-1713 (7.8%), E2-3-glucuronide (3.8%), and E2-3-sulfate (3.3%) (4). The assay employs an- tibody raised in New Zealand White rabbits after injection of E 1-glucuronide (E1G; Steraloids Inc., Wilton, NH) conjugated to bovine serum albumin (BSA). Antibody-coated microtiter plates were al- lowed to incubate at 4~ until the assay was per-

formed. Then 10 or 20 Ixl of control or diluted urine (20- to 50-fold) and E1G conjugate (horseradish per- oxidase conjugated with E1G via the mixed anhy- dride procedure) were added to the well to incubate for 2-8 hr, and the plates washed to separate the unbound portion. Substrate was then added and the enzyme conversion of the substrate was measured as a color change at 405 nm with an automatic mi- crotiter plate spectrophotometer. The absorbance of zero concentration (A ~ wells was used to deter- mine the maximum binding of enzyme conjugate to the antiserum. The absorbance values (A) of all standards, controls, and unknowns were divided by the average A ~ of each plate, and this ratio was multiplied by 100 to give the percentage of maxi- mum binding. Creatinine (Cr) concentrations were measured on all urine samples by a microtiter tech- nique (5) and the results are expressed as E1C per milligram of Cr. The intraassay coefficient of vari- ation for E1C was 11.2% and the interassay coeffi- cients of variation were 9.97 and 7.39% at approx- imately 35 and 65% binding on the standard curve, respectively. The interassay coefficient of variation for the creatinine assay at approximately 1 mg/ml was 2.41% (n -- 14).

Statistical Analyses

Statview software (Abacus Concepts , Inc. , Berkeley, CA) was used to calculated all statistics. Chi-square analysis was used to determine whether clinically relevant serum E2 values could be pre- dicted by a specific value of E l C. The least-squares method was used to determine the best-fit regres- sion line. Correlation analysis was performed using the Spearman rank method. A P value <0.05 was considered statistically significant on all analyses.

RESULTS

Mean E2 and E1C levels demonstrated a similar pattern on the days before hCG administration (Fig. 1). The log-transformed E2 and E1C data are pre- sented in Fig. 2. Within individual patients there was an excellent correlation between serum E2 and E1C/Cr, with 19 or 25 patients demonstrating cor- relation coefficients (R) >0.95. Based on the simi- larity of independent within-subject correlations, data from all subjects were grouped together for additional analysis. When this was done a strong correlation between urinary E1C/Cr and serum E2

Journal of Assisted Reproduction and Genetics, Vol. 11, No. 8, 1994

URINARY E1C IN IVF PATIENTS 407

E

2250

2000

1750

1500

1250

1000

750

500

250

0

Serum E2

-9 -7 -6 -5 -4 -3 -2 -1 0

c" 900 O

800 .

E 700

c 6OO

*~ 500

400

L~ 300

200

100

0 -9 -7 -6 -5 -4 -3 -2 -1 0

Day before hCG administration

Fig. 1. M e a n ( • serum E2 and E1C levels on days before

hCG administrat ion.

was observed (p -" 0.85, P < 0.0001). Further anal- ysis of the correlation of E2 and E1C/Cr by day from hCG injection revealed a bet ter correlation with lower values of E2. That is, the further from hCG injection, the bet ter the correlation between

q (,.) ,7,

.._1

3 5

3 2 5

3

2 7 5

2 5

2 2 5

2

1 7 5

1 .5

1 2 5

1

- x x x

~ x x xX ~,~ X tmk X

x x x x x x x x x x x x x

x ~ . ~ , ,'" ~" x- xX

x x x x x S xx: x , , i , i , i , i ,

6 1 8 2 2 2 2 4 2 6 2 8 3 3 . 2 3 4 3 6

Log E2

Fig. 2. Regress ion plot of log E2 and log E1C.

serum E2 and urinary E1G. The p correlation values (and P values) for days preceding hCG, with Day 0 being the day of hCG, were as follows: Day 0 = 0.65 (P = 0.001), Day - 1 = 0.59 (P = 0.013), Day - 2 = 0.62 (P = 0.01), Day - 3 = 0.81 (P = 0.005), Day - 4 = 0.91 (P = 0.017), and Day - 5 = 0.94 (P = 0.001).

Since EIC/Cr could lag behind serum E2, we also sought to find a correlation between E1C/Cr and the E2 value from the previous day but no improvement in the correlation coefficients was observed. Al- though there was a correlat ion between urinary E1C/Cr and serum E2, the strength of the associa- tion was not sufficient reliably to predict E2 by uri- nary E1C at higher values. Chi-square analysis failed to reveal a value of E1C that would ade- quately discriminate a value of E2/>2000 or i> 1500 pg/ml, a value clinically useful to predict ovarian hyperstimulation. However , the number of follicles greater than 12 mm showed an equally high corre- lation with E1C/Cr (p = 0.71, P < 0.0001) as it did with serum E2 (p = 0.74, P < 0.0001).

DISCUSSION

The present study reveals a similar rise in urinary E1C and serum E2 levels in patients undergoing ovarian hyperstimulation with hMG. Furthermore, a positive correlation between urinary E1G and se- rum E2 was demons t ra t ed . Individual pat ients tended to show the same pattern of both E2 and E1C levels. Our results are consistent with previous findings (4) in normally cycling women which dem- onstrated that the preovulatory rise in serum E2 was accompanied by a parallel rise in urinary E1C. However, in normally cycling (untreated) women, the rise in serum E2 and urinary E1C was consid- erably less than in treated women. E1C levels in untreated women were between 5 and 50 nmol/ mmol creatinine, in contrast to the values in this study of between 14 and 1775 nmol/mmol creati- nine.

This study also demonstrates that an association exists between E2 and E1C and that the correlation was strongest with lower values of E2. Others have also shown a similar association between serum E2 and urinary estrogen. Ishikawa and co-workers (6) showed a correlation coefficient of 0.80 between serum E2 and total urinary E2 as measured by latex agglutination inhibition reaction. Lessing and co- investigators (7) showed a progressively bet ter cor-

Journal o f Assisted Reproduction and Genetics, Vol. 11, No. 8, 1994

408 ALPER, HALVORSON, LASLEY, AND MORTOLA

relation (R = 0.37 to 0.86 from day 2 to day 8 of stimulation) between serum E2 and the log of uri- nary E l-glucuronide/Cr as measured by chemilumi- nescence immunoassay in women undergoing hMG therapy. Our study revealed the opposite, with bet- ter correlations with fewer days of stimulation, probably reflecting the differences in the assay per- formance.

The urinary estrogen metabolites in the normal mens t rua l cyc le are E l - g l u c u r o n i d e , E2-3- glucuronide, estriol-3-glucuronide, estradiol-17[3- glucuronide, and estriol-16a-glucuronide. While the E1C are the most abundant metabolites of circulat- ing E2 (8), other metabolites may provide a better predictor of increasing E2 production. Further- more, the pharmacologic effect of gonadotropins on estrogen metabolism and secretion is unknown and may account for the lack of a stronger correlation between E2 and E1C, especially among higher val- ues of serum E2.

We conclude that creatinine-corrected urinary E1C as measured by EIA provides an estimate of serum E2 levels in patients undergoing hMG ther- apy. Since none of the patients in this study devel- oped significant hyperstimulation, we were unable to determine the value of this test to predict ovarian hyperstimulation syndrome. The threshold of E1C that bests predicts the likelihood of ovarian hyper- stimulation awaits further clinical study. E1C data in patients with hyperstimulation syndrome would be of enormous value to determine whether our cur-

rent assay is able to predict the occurrence of the syndrome.

REFERENCES

1. Gelety TJ, Kerin JF, Surrey ES: Ovarian hyperstimulation syndrome: Diagnosis and treatment. Infertil Reprod Med 1992 ;3:795- 809

2. Lasley BL, Stabenfeldt GH, Overstreet JW, Hansen FW, Czekala N, Munro C J: Urinary hormone levels at the time of ovulation and implantation. Fertil Steril 1985;43:861-867

3. Czekala NM, Overstreet JW, Hanson FW, Stabenfeldt GH, Lasley BL: Assessment of follicular function in women by measurement of urinary estrone conjugates. Fertil Steril 1986;46:604-609

4. Munro C J, Stabenfeldt GH, Cragun JR, Addiego LA, Over- street JW, Lasley BL: Relationship of serum estradiol and progesterone concentrations to the excretion profiles of their major urinary metabolites as measured by enzyme immuno- assay and radioimmunoassay. Clin Chem 1991;37:838-844

5. Taussky HH: A microcolorimetric determination of creati- nine in urine by the Jaffe reaction. J Biochem 1954;208:853- 861

6. Ishikawa M, Hoshiai H, Tozawa H, Fukaya T, Yajima A: Monitoring follicular maturation through measurement of urinary estrogen excretion by latex agglutination inhibition reaction. Fertil Steril 1987;48:688-690

7. Lessing JB, Peyser MR, Gilad S, Amit A, Kogosowski A, Yovel I, Barak Y, David MP: Es t rone-3-g lucuronide chemiluminescence immunoassay: An alternative method for monitoring induction of ovulation with human meno- pausal gonadotropin in an in vitro fertilization program. Fer- til Steril 1987;48:450-453

8. Baker TS, Jennison KM, Kellie AE: The direct radioimmu- noassay of glucuronides in human female urine. Biochem J 1979;177:729-738

Journal of Assisted Reproduction and Genetics, Vol. 11, No. 8, 1994