Evaluation of basal estradiol levels in assisted reproductive technology cycles

FERTILITY AND STERILITY VOL. 74, NO. 3, SEPTEMBER 2000 Copyright 2000 American Society for Reproductive Medicine Published by Elsevier Science Inc. Printed on acid-free paper in U.S.A. Evaluation of basal estradiol levels in assisted reproductive technology cycles John L. Frattarelli, M.D., a Paul A. Bergh, M.D., b,c Michael R. Drews, M.D., b,c Fady I. Sharara, M.D., d and Richard T. Scott, Jr., M.D. b,c Combined Federal Program for Reproductive Endocrinology at Walter Reed Army Medical Center, National Naval Medical Center, Uniformed Services University for the Health Sciences, and National Institutes of Health, Bethesda, Maryland; The Institute for Reproductive Medicine and Science of Saint Barnabas Medical Center, Livingston, New Jersey; Reproductive Medicine Associates of New Jersey, Morristown, New Jersey; and University of Maryland, Baltimore, Maryland Received December 13, 1999; revised and accepted March 21, 2000. Reprint requests and present address: John L. Frattarelli, M.D., Department of Obstetrics and Gynecology, Tripler Army Medical Center, 1 Jarrett White Road, Tripler AMC, HI 76859-5000 (FAX: 808-433-1552; E-mail: jfratt@yahoo.com). a Combined Federal Program for Reproductive Endocrinology at WRAMC, NNMC, USUHS, and NIH. b The Institute for Reproductive Medicine and Science of Saint Barnabas Medical Center. c Current address: Reproductive Medicine Associates of New Jersey. d Department of Obstetrics, Gynecology, and Reproductive Sciences, University of Maryland. The opinions contained herein are the private views of the authors and are not to be construed as official or as reflecting the views of the Department of the Army or the Department of Defense. 0015-0282/00/$20.00 PII S0015-0282(00)00693-2 Objective: To determine if basal E 2 screening increases the diagnostic accuracy of basal FSH screening and to determine whether basal E 2 levels correlate with outcome in ART cycles. Design: Retrospective. Setting: Tertiary care center. Patient(s): Two thousand six hundred thirty-four infertility patients. Intervention(s): Cycle outcome was evaluated after grouping patients by basal E 2 levels beginning at 20 pg/ml and extending to 100 pg/ml at 10 pg/ml increments. Main Outcome Measure(s): Retrieved oocytes, pregnancy rate, and cancellation rate. Result(s): Cancellation rates were significantly increased in patients with basal E 2 levels of 20 pg/ml or 80 pg/ml. Basal E 2 levels neither predicted pregnancy outcome nor correlated with ovarian response in those patients not canceled. Conclusion(s): Patients with basal E 2 levels of 20 pg/ml or 80 pg/ml had an increased risk for cancellation. Basal E 2 was predictive of stimulation parameters in patients 40 years or older. For those patients who proceeded to retrieval, there were no differences in pregnancy or delivery rates relative to basal E 2 levels. This suggests that irrespective of basal E 2 levels patients who produce more than three maturing follicles in response to stimulation have adequate ovarian reserve as evidenced by their pregnancy rates. (Fertil Steril 2000;74:518 24. 2000 by American Society for Reproductive Medicine.) Key Words: Assisted reproductive technology, in vitro fertilization, estradiol, follicle-stimulating hormone, ovarian reserve, pregnancy rates, cancellation rates, review The evaluation of ovarian reserve has been the focus of much clinical research over the last several years (1 15). A number of tests have been suggested to prognosticate ovarian responsiveness to exogenous gonadotropin stimulation, the quality of the oocytes retrieved, and the subsequent implantation and pregnancy rates (1 13). Most of these tests are based on circulating FSH levels in the basal state (basal day 3 FSH levels) or after treatment with the antiestrogen clomiphene citrate (the clomiphene citrate challenge test). The prognostic value of these tests has been clearly shown by a number of investigators in a wide variety of settings (1 13). Despite their increasing acceptance in clinical practice, the tests continue to have a variety of significant shortcomings. The principal limitation is the lack of predictive value of a normal result. Although the predictive value of an abnormal result is high with pregnancy rates of 5% (in many cases 1%), the predictive value of a normal result is limited because many patients with poor gonadotropin responsiveness and low-quality oocytes and embryos have normal screening results (11). Thus, the available tests for diminished ovarian reserve are specific but not sensitive. The lack of sensitivity has led a number of investigators to evaluate other tests to identify those patients whose ovarian reserve is insufficient to allow them to conceive (16 21). Recently, serum basal E 2 has been evaluated (17 20). This particular hormone is most attractive 518

because it is a direct product of the ovary, which may be assayed in virtually any endocrine laboratory. It has been postulated that the elevation in basal E 2 levels may suppress otherwise elevated FSH levels into the normal range, thus masking what might have been an abnormal test result. Three groups of investigators described decreases in clinical outcomes for patients with elevated basal E 2 levels (17 19). The purposes of this study were to evaluate the predictive value of basal E 2 levels in assisted reproductive technology (ART) cycles based on a large cohort of patients, to compare basal E 2 levels with clinical outcomes of pregnancy and cancellation rates, and to determine if specific basal E 2 levels predicted clinical response. MATERIALS AND METHODS Population All patients who initiated IVF cycles in our ART program between June 1996 and November 1998 (n 3,365) and whose basal day 3 FSH and E 2 levels were determined in our laboratory were considered for inclusion in the study. Patients were included independent of their age, diagnoses, or reproductive history. For patients (n 482) undergoing multiple cycles (n 731), only their first cycle was included for evaluation. Patients (n 21) with elevated FSH levels ( 14 miu/ml) were excluded from further analysis because their poor prognosis was previously established (7). Likewise, patients (n 137) who had ovarian screening at an outside laboratory were excluded; remaining were 2,476 patients who met the inclusion criteria. Institutional Review Board approval from Saint Barnabas Medical Center was obtained. Experimental Design All day 3 basal FSH, LH, and E 2 measurements were made during a spontaneous cycle within 6 months before undergoing ART. Patients underwent stimulation with a GnRH analogue (GnRH-a) down-regulation followed by stimulation with exogenous gonadotropins or concomitant initiation of GnRH-a and exogenous gonadotropins ( flare cycles), as previously described (21, 22). Specifically, those patients (n 260) who showed poor follicular response in prior ART cycles and those patients thought to be at risk for poor response based on their age and hormone values were started on a follicular phase GnRH-a flare protocol. Retrieval, transfer, and all andrological and embryological procedures were completed as previously described (23, 24). Biochemical pregnancies were defined as having a positive pregnancy test on luteal day 16 with a rising titer confirmed by a second hcg level but with loss of the pregnancy before the development of sonographic evidence of the pregnancy. Clinical pregnancy was defined as a pregnancy documented by sonographic confirmation of an intrauterine pregnancy. Ongoing or delivered pregnancies were defined as gestations that progressed beyond 13 weeks of gestational age that were proceeding normally or that had been delivered. Cycles were canceled for failing to produce three or more expanding follicles and/or failure to respond to gonadotropins with an adequate E 2 response of 500 pg/ml. Laboratory Assays The LH, FSH, and E 2 assays were all performed with the Immulite immunoassay system (Diagnostic Products Corporation, Los Angeles, CA). The interassay and intraassay coefficients of variation were 6.6% and 5.4% for FSH, 5.8% and 4.6% for LH, and 5.4% and 4.4% for E 2, respectively. Statistical Analysis Patients with normal FSH levels were grouped by basal E 2 levels between 20 pg/ml and 100 pg/ml at 10 pg/ml increments. Cancellation, biochemical pregnancy, clinical pregnancy, and ongoing or delivery rates were calculated for each increment. Pregnancy and cancellation rates above each threshold were evaluated with receiver operating characteristic (ROC) curves to determine if there were any obvious break points above which there was an obvious change in these rates. Contingency table analysis was then used to assess the pregnancy and cancellation rates above and below the selected threshold value. Spearman s correlation coefficients were calculated to measure the strength of the association between the ovarian response and stimulation parameters. The predictive accuracy of these variables was compared by simple linear regressions where the ovarian response was represented by the number of oocytes retrieved. Multilinear regression was then used to determine the best predictors of outcome. A Kruskal-Wallis one-way analysis of variance was used to compare basal E 2 levels, patient age, day 3 FSH levels, and day 3 LH levels within each of the pregnancy groups and the cancellation group to determine if there were differences in the population means. For pairwise multiple comparison, Dunn s method was used. A Mann-Whitney rank sum test was used to compare means between two groups. An alpha error of 0.05 was considered significant for all comparisons. All data are reported as means with their associated standard deviations. RESULTS A total of 3,365 ART cycles were initiated during the study interval. There were 482 patients who underwent a total of 731 repeat cycles. These 731 repeat cycles were excluded from our analysis. Ovarian reserve screening was determined in our laboratory in 2,497 of the remaining 2,634 patients. Of these, 2,476 were performed in women with normal basal FSH levels ( 14 miu/ml). This group (n 2,476) constituted the population for this study. Cancellation at any time after initiating gonadotropins occurred in 343 patients, and 2,133 patients went to oocyte retrieval. There FERTILITY & STERILITY 519

TABLE 1 Demographics and stimulation characteristics for the 2,476 ART patients in the study group. Variable No. Age Basal FSH level Basal E 2 level Basal LH level Peak E 2 level No. of oocytes No. of days of gonadotropins No. of ampules of gonadotropins Total population 2,476 35.6 3.9 4.92 2.2 41.2 27.2 4.38 3.1 2,496 1,305 16.2 8.4 8.7 1.3 42.9 19.9 Canceled cycles 343 37.2 3.8 4.63 3.0 43.3 47.0 3.44 2.5 Not pregnant 712 35.8 4.1 4.97 2.1 40.9 22.2 4.39 3.8 2,453 1,405 15.0 8.3 8.8 1.3 43.9 12.8 Total pregnancies 1,421 35.1 3.8 4.97 2.0 40.8 23.6 4.60 2.7 2,517 1,251 16.8 8.4 8.7 1.3 42.5 14.7 Ongoing/delivered 1,131 34.8 3.6 a 4.99 2.0 41.0 23.9 4.71 2.8 a 2,530 1,265 17.0 8.3 a 8.7 1.3 41.2 13.0 Clinical loss 149 36.2 4.0 4.95 2.1 39.5 22.3 4.22 2.1 2,543 1,259 15.7 8.2 8.8 1.3 44.3 16.8 Biochemical pregnancies 141 36.2 4.0 4.91 2.3 41.3 22.5 4.10 2.2 2,394 1,134 16.3 8.7 8.8 1.3 43.8 13.0 Note: Values are means SD. a Significantly different from the not pregnant, clinical pregnancy loss, and biochemical pregnancy groups using a Kruskal-Wallis one-way analysis of variance (P.01). were 1,421 pregnancies for a crude pregnancy rate of 66.6% per retrieval. Of these, 141 were biochemical (9.9% of pregnancies) and 149 were clinical losses (10.5% of pregnancies). There were a total of 1,131 ongoing pregnancies and deliveries (79.6% of pregnancies; 53% per retrieval). Table 1 displays the demographics and stimulation characteristics for the study group. With use of a Kruskal-Wallis one-way analysis of variance, there were no statistically significant differences in the mean basal E 2 levels (P.07) between those women who did not conceive (40.9 22.2 pg/ml), those who had biochemical pregnancies (41.3 22.5 pg/ml), those who had clinical losses (39.5 22.3 pg/ml), or those with ongoing pregnancies or deliveries (41.0 23.9 pg/ml). Similarly, there were no statistically significant differences in basal day 3 FSH levels (P.77), stimulation quality as evidenced by peak E 2 levels (P.70), days of stimulation (P.21), or ampules of gonadotropins administered (P.14). Patients with ongoing pregnancies or deliveries had significantly more oocytes retrieved (17.0 8.3) than did patients not conceiving (15.0 8.3), patients experiencing a biochemical loss (16.3 8.7), and patients experiencing a clinical loss (15.7 8.2) (P.01). Patients with ongoing pregnancies or deliveries were significantly younger (34.8 3.6) than patients not conceiving (35.8 4.1), patients experiencing a biochemical loss (36.2 4.0), and patients experiencing a clinical loss (36.2 4.0) (P.01). Likewise, patients with ongoing pregnancies or deliveries had higher day 3 LH levels (4.71 2.8 miu/ml) than patients not conceiving (4.39 3.8 miu/ml), patients experiencing a biochemical loss (4.11 2.2 miu/ml), and patients experiencing a clinical loss (4.22 2.1 miu/ml) (P.01) (Table 1). When the stimulation characteristics of peak E 2 and number of oocytes retrieved were evaluated throughout the full range of E 2 values with efficiency curves, no obvious difference was seen until the groups were subdivided by age. The only difference noted was in the patients 40 years. In these patients, an E 2 level of 80 pg/ml was predictive of a poor response producing a lower number of oocytes (14.34 8.04 vs. 9.15 4.38) (P.01) and a lower peak E 2 level (2,278 1,157 vs. 1,439 403) (P.01). In this group of patients, the number of days of stimulation (8.9 1.3 vs. 8.8 1.2) and ampules of gonadotropins used (49.2 10.1 vs. 52.8 5.7) were not altered significantly by an elevation in E 2 level. The overall pregnancy and cancellation rates relative to basal E 2 measurements are shown in Figure 1. Contingency FIGURE 1 Overall pregnancy and cancellation rates for each basal E 2 range. *Contingency table analysis, P.01. 520 Frattarelli et al. Basal estradiol levels in ART cycles Vol. 74, No. 3, September 2000

TABLE 2 Distribution of ART cycles and their outcome relative to basal E 2 levels among the patient population. Basal E 2 level No. of retrievals No. of total pregnancies (%) No. of ongoing/ delivered (%) No. of clinical losses (%) No. of biochemical pregnancies (%) No. of canceled cycles (%) 20 267 167 (62.5) 139 (52.1) 16 (9.6) 12 (7.2) 83 (23.7) a 20 29 395 274 (69.4) 207 (52.4) 32 (11.7) 32 (11.7) 65 (14.1) 30 39 525 355 (67.6) 275 (52.4) 48 (13.5) 32 (9.0) 77 (12.8) 40 49 445 290 (65.1) 240 (53.9) 24 (8.3) 27 (9.3) 44 (9.0) 50 59 239 165 (69.0) 128 (53.6) 16 (9.7) 22 (13.3) 26 (9.8) 60 69 133 84 (63.2) 73 (54.9) 4 (4.8) 7 (8.3) 18 (11.9) 70 79 59 41 (69.5) 33 (55.9) 4 (9.8) 4 (9.8) 8 (11.9) 80 89 26 17 (61.5) 14 (53.8) 1 (5.8) 2 (11.8) 6 (18.8) a 90 99 10 9 (90.0) 8 (80.0) 1 (11.1) 0 (0) 4 (28.6) 100 34 20 (58.8) 14 (41.2) 3 (15.0) 3 (15.0) 12 (26.1) Total 2,133 1,421 (66.6) 1,131 (53.0) 149 (10.5) 141 (9.9) 343 (13.9) Note: Values represent the total number of patients per group with percentages in parentheses. a Represents a break point in the data (P.01). table analyses detected no differences in overall pregnancy rates for any basal E 2 level (Table 2). The ROC curves for overall pregnancy rates or pregnancy loss rates revealed no obvious differences throughout the full range of values. Similar evaluations for the subset of patients 35 years of age, 35 39.9 years of age, and those 40 years of age revealed no differences in pregnancy rates or outcome. Thus, basal E 2 levels did not correlate with pregnancy rates or cycle outcome. The efficiency curves for cancellation rates suggested an increased risk for patients whose basal E 2 level was 20 ng/ml or 80 pg/ml. Contingency table analyses confirmed that the risk for cancellation was increased in patients with basal E 2 levels of 20 ng/ml (P.01, odds ratio [OR] 2.23, 95% confidence interval [CI] 1.6 2.4) or of 80 pg/ml (P.01, OR 2.02, 95% CI 1.4 2.9) (Table 2). For E 2 levels of 20 ng/ml, the positive predictive value was 23.7%, and the negative predictive value was 87.8%, whereas for E 2 levels of 80 pg/ml the positive predictive value was 23.9%, and the negative predictive value was 86.5%. Evaluating the age-related risk of cancellation offered no additional statistical benefit. A linear regression analysis was used to determine the correlation of patient age, basal FSH, basal LH, and basal E 2 levels with ovarian response (i.e., number of oocytes retrieved). Patient age (r 0.36, P.01) and basal FSH level (r 0.82, P.01) revealed a significant negative linear correlation with ovarian response. Basal LH level showed a positive linear correlation with ovarian response (r.37, P.01). Basal E 2 level did not correlate with ovarian response (P.59). Likewise, a multiple linear regression analysis revealed that basal FSH level, basal LH level, and patient s age were the best predictors of ovarian response. With use of a Mann-Whitney rank sum test, the canceled patients were compared with those patients not canceled. Canceled patients had significantly higher basal E 2 levels (43.3 47 pg/ml) than noncanceled patients (40.9 23 pg/ml) (P.01). Likewise, canceled patients were significantly older (P.01) and had lower basal FSH (P.03) and LH levels (P.01). Similar evaluations were performed on subsets of patients who were canceled vs. not canceled, subdividing the patients into the following groups: 35 years of age, 35 39.9 years of age, and 40 years of age. Basal serum LH levels were significantly lower in all of the canceled patients (P.01). After subdividing by age, basal serum FSH levels did not statistically significantly differ in canceled vs. noncanceled patients. It is of interest that the basal E 2 levels for canceled patients were significantly lower in the group 35 years of age (P.01) and significantly higher in the two groups older than 35 years of age (P.01) (Table 3). DISCUSSION The data in this study indicated that undetectable or elevated E 2 levels on cycle day 3 correlate with higher cancellation rates during ART cycles. However, if patients continued to retrieval, there was no difference in the pregnancy or delivery rates. The E 2 levels did not correlate with stimulation parameters except in the patients 40 years and older, where a lower peak E 2 and fewer oocytes were seen with an E 2 level of 80 pg/ml. Ours is the first study that documented a detrimental effect of a low E 2 value on patients undergoing ART. We found that in the 350 patients with an E 2 level of 20 pg/ml there was a significant increase in cancellation rates. The theoretical principle that supports basal E 2 screening FERTILITY & STERILITY 521

TABLE 3 Canceled patients vs. noncanceled patients: comparison of the overall patient population and the three age subsets ( 35, 35 39, 40 years). No. Age (y) FSH LH E 2 Canceled (all) 343 37.2 3.8 4.63 3.0 3.44 2.5 43.3 47 Not canceled (all) 2,133 35.3 3.9 4.97 2.0 4.53 3.1 40.8 23 P value 0.01 a 0.03 a 0.01 a 0.01 a Canceled ( 35 years) 93 32.2 2.5 4.72 2.9 3.98 2.8 35.2 18 Not canceled ( 35 years) 993 32.0 2.3 4.87 2.0 4.61 2.4 39.6 24 P value 0.17 0.48 0.01 a 0.01 a Canceled (35 39.9 y) 157 37.7 1.4 4.69 3.2 3.21 2.2 46.2 58 Not canceled (35 39.9 y) 859 37.3 1.5 5.06 2.0 4.54 3.9 42.4 23 P value 0.01 a 0.12 0.01 a 0.01 a Canceled ( 40 y) 93 41.4 0.9 4.45 2.7 3.28 2.6 46.5 47 Not canceled ( 40 y) 281 41.3 0.9 5.04 2.2 4.23 2.7 40.7 22 P value 0.11 0.15 0.01 a 0.55 Note: Values are means SD. a Values represent statistically significant difference (P.05) using a Mann Whitney rank sum test. is the ability to detect patients with shortened follicular phases who may have progressed far enough into their follicular phases to invalidate the evaluation of their basal FSH levels. This might be thought of in simpler terms as a suppression of an otherwise abnormal FSH level into the normal range by elevated E 2 levels. However, this suppression is more likely the impact of a number of secretory products and not E 2 alone. Evaluation of other secretory products of the ovary are also currently being evaluated. Recent reports suggest that measurement of dimeric inhibin levels may add additional information to the more traditional forms of ovarian reserve screening (16). Basal FSH screening is only valid during a relatively narrow window within a given menstrual cycle. It has been shown that the age-related elevation in FSH levels occurs exclusively in the early follicular phase and that any differences become undetectable by the mid to late follicular phase (25). This is likely due to the increased secretion of E 2 and a variety of ovarian peptides whose effects are integrated to lower FSH levels. For these reasons, there has been a long-standing concern that the sensitivity of FSH screening might not be valid in patients with E 2 levels that appear to be beyond those typically seen on day 3. Consistent with that concern is the fact that patients with elevated progesterone levels on day 10 during the clomiphene citrate challenge test have diminished ovarian reserve even if their FSH levels are normal (26). This evidence of luteinization indicates a markedly shortened follicular phase, which is one of the hallmarks of reproductive aging. These patients do uniformly poorly despite having FSH values in the normal range. Table 4 summarizes the four prior publications detailing the value of basal E 2 measurements as a component of ovarian reserve screening. The first, by Licciardi et al. (17), described a decrease in oocyte number and pregnancy rates with increasing E 2 levels, which were measured during an ART cycle not using a GnRH-a. However, the data as presented in the manuscript, indicated that once the investigators controlled for elevated FSH levels there was no statistically significant difference in pregnancy rates. The data by Licciardi et al. (17) seemed to indicate that basal E 2 screening did not add significant information beyond that obtained with basal FSH screening alone. Smotrich et al. (18) found a significant decline in pregnancy rates and increase in cancellation rates among 292 patient cycles with rising basal E 2 levels, which were measured in a cycle proximate to the stimulation cycle. This correlation persisted even after controlling for basal FSH status and patient age. Smotrich et al. (18) described a clinically meaningful threshold level at approximately 80 pg/ml based on 18 patients (27 cycles) with estradiol levels above this threshold. Evers et al. (19) measured E 2 levels on cycle day 3 in a spontaneous cycle before beginning GnRH-a; they evaluated 231 cycles and found an E 2 threshold level of 60 pg/ml where they obtained no pregnancies. Likewise, in their 16 patients with E 2 levels of 60 pg/ml, Evers et al. (19) also found a significantly lower number of oocytes retrieved. In contrast to the three above-mentioned studies, Vazquez et al. (20) used a study population of 248 patient cycles in which E 2 levels were evaluated on cycle day 2 with GnRH-a and gonadotropins added thereafter; they concluded that E 2 levels below the age of 35 years were not helpful. Common themes in the previous publications evaluating 522 Frattarelli et al. Basal estradiol levels in ART cycles Vol. 74, No. 3, September 2000

TABLE 4 Comparison of previously published studies evaluating the prognostic value of E 2 in patients undergoing ART cycles to the current study. Study ART cycles a E 2 threshold ART cycles above threshold b Results Licciardi et al., 1995 592 30 332 Decreased pregnancy rates 60 86 Decreased oocyte number 75 30 No pregnancies Smotrich et al., 1995 292 80 27 Increased cancellation rates, decreased pregnancy rates 100 15 No pregnancies Evers et al., 1998 231 60 16 Decreased oocyte number, increased cancellation rates, and no pregnancies Vazquez et al., 1998 248 25? Decreased implantation and pregnancy rates only in patients older than 35 years Frattarelli et al., 2000 2,476 20 350 Increased cancellation rates 80 92 Increased cancellation rates a Total number of ART cycles evaluated in the published article. b Total number of ART cycles above the published threshold E 2 level. basal E 2 levels and ovarian response in ART were small patient groups and E 2 values not based on statistical cutoff points (Table 4). In the article by Licciardi et al. (17), 30 patients had E 2 levels of 75 pg/ml, and none of the patients became pregnant. Evers et al. (19) had 16 patients with E 2 levels of 60 pg/ml, which resulted in no pregnancies. Likewise, Smotrich et al. (18) had 27 patient cycles with E 2 levels of 80 pg/ml resulting in 3 pregnancies and 15 patient cycles with E 2 levels of 100 pg/ml, which resulted in no pregnancies. In contrast, we had 92 patients with basal E 2 levels of 80 pg/ml, resulting in 46 pregnancies (65.7%) and 22 cancellations (33.9%) (Table 2, Fig. 1). Furthermore, we had 46 patients with basal E 2 levels of 100 pg/ml, resulting in 20 pregnancies (58.8%) and 12 cancellations (26.1%). With a sample size of 2,476 patients, the current study did detect a difference in reproductive performance in patients with undetectable or elevated E 2 levels on cycle day 3. These patients were more likely to respond poorly to gonadotropin stimulation, producing three or fewer follicles. However, among most patients who did respond well enough to stimulation to proceed to retrieval, the current study could not identify a basal E 2 value at which the probability for pregnancy declined. These data suggest that some patients with normal ovarian reserve will occasionally have a delayed or an accelerated cycle with either low or midfollicular levels of E 2 being present by day 3. Thus, the population of patients who truly have diminished ovarian reserve may be diluted by others having an occasional accelerated cycle. This is consistent with the data found in the original description of basal FSH screening when E 2 levels did not seem to add significantly to the test. In our observations, we noted some interesting results (Table 3). First, after excluding all patients with elevated day 3 FSH levels, the patients who were canceled had a significantly lower basal FSH level. Although not statistically significant, this finding persisted when the patients were subdivided by age. Second, the canceled patients had significantly higher basal E 2 levels. However, when subdivided by age, the canceled patients 35 years of age had a significantly lower basal E 2 level, whereas the canceled patients 35 years of age had higher basal E 2 levels than those patients not canceled. This finding could suggest that diminished ovarian reserve in younger patients may be manifested by a delayed E 2 response, whereas in older patients an accelerated response is seen. As with any retrospective study, introduction of bias into the study must be evaluated. We tried to minimize bias by including all patients meeting our inclusion criteria regardless of diagnosis and age. Subsequently, we statistically evaluated each of the subgroups by age. Theoretically, including patients with the diagnosis of PCOS could bias our results by increasing our total number of oocytes retrieved and elevating the pregnancy rate in those patients with an elevated E 2. However, only 3% of our patients had polycystic ovary syndrome (PCOS) as their primary diagnosis. When those patients with PCOS were removed from the analysis, there was no change in any of our statistical values. Therefore, we retained the patients with PCOS in our analysis. The inclusion of patients undergoing a microdose flare protocol could also have introduced bias into our data. These 260 patients on the microdose protocol were not statistically older than the remaining population of 2,216. Because by removing these 260 patients from the data analysis no change in the statistical values occurred, these patients also were retained in the data analysis. In summary, the addition of basal E 2 measurements to ovarian reserve screening may allow clinicians to identify patients who are at increased risk for cancellation. In those FERTILITY & STERILITY 523

patients 40 years, an elevated E 2 level of 80 pg/ml correlated with poor stimulation outcome as evidenced by a lower peak E 2 level and fewer oocytes. Nevertheless, most patients stimulated adequately, proceeded to retrieval, and then achieved pregnancy rates commensurate with their age. Patients with basal E 2 levels that are undetectable or above the normal early follicular range should not be counseled that they have diminished ovarian reserve. Although cycle outcome was poorer, elevated basal E 2 levels would not seem to be a reason to cancel or postpone a patient s stimulation cycles. References 1. Scott RT, Hofmann GE, Oehninger S, Muasher SJ. Intercycle variability of day 3 follicle-stimulating hormone levels and its effect on stimulation quality in in vitro fertilization. Fertil Steril 1990;53:297 302. 2. Scott RT, Toner JF, Muasher SJ, Oehninger SC, Robinson S, Rosenwaks Z. Follicle-stimulating hormone levels on cycle day 3 are predictive of in vitro fertilization outcome. Fertil Steril 1989;51:651 4. 3. Navot D, Rosenwaks Z, Margalioth EJ. Prognostic assessment of female fecundity. Lancet 1987;2:645 7. 4. Hofmann GE, Sosnowski J, Scott RT, Thie J. Efficacy of selection criteria for ovarian reserve screening using the clomiphene citrate challenge test in a tertiary fertility center population. Fertil Steril 1996;66:49 53. 5. Loumaye E, Billion JM, Mine JM, Psalit I, Pensis M, Thomas K. Prediction of individual response to controlled ovarian hyperstimulation by means of a clomiphene citrate challenge test. Fertil Steril 1990;53:295 301. 6. Pearlstone AC, Fournet N, Gambone JC, Pang SC, Buyalos RP. Ovulation induction in women age 40 and older: the importance of basal follicle-stimulating hormone level and chronological age. Fertil Steril 1992;58:674 9. 7. Scott RT, Toner JF, Muasher SJ, Oehninger SC, Robinson S, Rosenwaks Z. Follicle-stimulating hormone levels on cycle day 3 are predictive of in vitro fertilization outcome. Fertil Steril 1989;51:651 4. 8. Scott RT, Leonardi MR, Hofmann GE, Illions EH, Neal GS, Navot D. A prospective evaluation of clomiphene citrate challenge test screening in the general infertility population. Obstet Gynecol 1993;82:539 45. 9. Scott RT, Dellinger CL, Leonardi MR, Illions EH, Hofmann GE, Navot D. Evaluation of the significance of a poor estradiol response during the clomiphene citrate challenge test. Fertil Steril 1993;60:242 6. 10. Scott RT, Opsahl MS, Leonardi MR, Neal GS, Illions EH, Navot D. Life table analysis of pregnancy rates in a general infertility population relative to ovarian reserve and patient age. Hum Reprod 1995;10:1706 10. 11. Scott RT, Hofmann GE. Prognostic assessment of ovarian reserve. Fertil Steril 1995;63:1 11. 12. Tanbo T, Dale PO, Abyhom T, Stokke KT. Follicle-stimulating hormone as a prognostic indicator in clomiphene citrate/human menopausal gonadotrophin-stimulated cycles for in vitro fertilization. Hum Reprod 1989;6:647 50. 13. Toner JP, Philput CB, Jones GS, Muasher SJ. Basal follicle-stimulating hormone level is a better predictor of in vitro fertilization performance than age. Fertil Steril 1991;55:784 91. 14. Klein NA, Battaglia DE, Fujimoto VY, Davis GS, Bremner WJ, Soules MR. Reproductive aging: accelerated ovarian follicular development associated with a monotropic follicle-stimulating hormone rise in normal older women. J Clin Endocrinol Metab 1996;81:1038 45. 15. Klein NA, Battaglia DE, Clifton DK, Bremner WJ, Soules MR. The gonadotropin secretion pattern in normal women of advanced reproductive age in relation to the monotropic FSH rise. J Soc Gynecol Invest 1996;3:27 32. 16. Seifer DB, Gardiner AC, Lambert-Messerlian G, Schneyer AL. Differential secretion of dimeric inhibin in cultured luteinized granulosa cells as a function of ovarian reserve. J Clin Endocrinol Metab 1996;81: 736 9. 17. Licciardi FL, Liu HC, Rosenwaks Z. Day 3 estradiol serum concentrations as prognosticators of ovarian stimulation response and pregnancy outcome in patients undergoing in vitro fertilization. Fertil Steril 1995; 64:991 4. 18. Smotrich DB, Widra EA, Gindoff PR, Levy MJ, Hall JL, Stillman RJ. Prognostic value of day 3 estradiol on in vitro fertilization outcome. Fertil Steril 1995;64:1136 40. 19. Evers JLH, Slaats P, Land JA, Dumoulin JCM, Dunselman GAJ. Elevated levels of basal estradiol-17 predict poor response in patients with normal basal levels of follicle-stimulating hormone undergoing in vitro fertilization. Fertil Steril 1998;69:1010 4. 20. Vazquez ME, Verez JR, Stern JJ, Najar AG, Asch RH. Elevated basal estradiol levels have no negative prognosis in young women undergoing ART cycles. Gynecol Endocrinol 1998;12:155 9. 21. Droesch K, Muasher ST, Brzyski R, Jones SG, Simonetti S, Liu HCH, et al. Value of suppression with a GnRH-a prior to gonadotropin stimulation for in vitro fertilization. Fertil Steril 1989;51:292 5. 22. Scott RT, Navot D. Enhancement of ovarian responsiveness with microdoses of gonadotropin-releasing hormone agonist during ovulation induction for in vitro fertilization. Fertil Steril 1994;61:880 5. 23. Cohen J, Alikani M, Trowbridge J, Rosenwaks Z. Implantation enhancement by selective assisted hatching using zona drilling of embryos with poor prognosis. Hum Reprod 1992;7:685 91. 24. Schoolcraft WB, Schlenker T, Gee M, Jones GS, Jones HW. Efficacy of assisted hatching in poor prognosis IVF candidates. Fertil Steril 1994; 62:551 4. 25. Lenton EA, Sexton L, Lee S, Cooke ID. Progressive changes in LH and FSH and LH:FSH ratio in women throughout reproductive life. Maturitas 1988;10:35 43. 26. Hofmann GE, Scott RT Jr, Horowitz GM, Thie J, Navot D. Evaluation of the reproductive performance of women with elevated day 10 progesterone levels during ovarian reserve screening. Fertil Steril 1995;63:979 83. 524 Frattarelli et al. Basal estradiol levels in ART cycles Vol. 74, No. 3, September 2000