Accuracy of ovarian reserve tests

Transcription

1 Human Reproduction vol. no. pp.-6, Accuracy of ovarian reserve tests BUlent Gülekli ' ', Yesim Bulbul, Ata Onvural, Kutsal Yorukoglu, Cemal Posaci, Namik Demir and Oktay Erten,, To whom correspondence should be addressed at: McGill Reproductive Center, Department of Obstetrics and Gynecology, Royal Victoria Hospital, 6 Pine Avenue West, # F6., Montreal, Quebec Canada HA A Several tests predict ovarian reserve in women undergoing assisted reproductive technologies. However, the accuracy of these tests in assessing the number of the remaining follicles within the ovary (ovarian reserve) has not been previously validated. The aim of this study was to assess the accuracy of ovarian reserve tests, namely basal and clomiphene-stimulated follicle stimulating hormone () concentrations and gonadotrophin-releasing hormone (GnRH) agonist stimulation test in predicting the number of the follicles within the ovaries. The ovaries of parous women over years of age who underwent oophorectomy were examined histologically for follicle number. Early follicular phase serum, clomiphene citrate challenge tests (CCCT) and GnRH agonist stimulation test (GAST) were performed in the menstrual cycle prior to the surgery. The predictive value of these tests was then assessed. A positive correlation was detected between basal serum oestradiol concentrations and follicles per unit tissue but no significant correlation was detected between basal and clomiphene-stimulated and follicles per unit tissue. The receiver operator characteristic curves indicated that the clomiphene citrate challenge test was the most accurate of the three tests assessed. In conclusion, none of the tests in this study accurately reflects ovarian reserve. Key words: ageing/ovarian reserve tests/ovary Introduction The current trend towards delayed childbearing has decreased the reproductive potential of women. This diminished fertility has been expressed as the decreased number and quality of oocytes, changes in fertilizability and implantation, and increased risk of embryonic chromosomal abnormalities. Female fecundity is related to the total number of primordial follicles remaining within the ovaries, referred to as ovarian reserve, and this declines with age (Faddy and Gosden, 6). The limited predictive value of age alone in estimating Attempts have been made to use follicular density in ovarian biopsy specimens as a direct means of assessing ovarian reserve (Lass et al., ). Although this may be a novel clinical tool for the evaluation of ovarian reserve, the limitation of this test is that follicular density within the biopsy specimen may not accurately represent the density of follicles within the whole ovary. The purpose of this study, therefore, was to examine the accuracy of the ovarian reserve tests by comparing the ability of basal serum, CCCT, and GAST to predict the number of follicles within the ovaries as assessed by histology, in women over years of age who underwent oophorectomy. Materials and methods This study was performed in the Obstetrics and Gynaecology Depart ment of Dokuz Eylul University School of Medicine, Izmir, Turkey, between February 6 and May. The study was approved by the Human Investigation Review Committee of the Department, and informed written consent was obtained from all the participants. Twenty-four pre-menopausal parous women over years of age, with regular menses, and about to undergo laparotomy with unilateral or bilateral salpingo-oophorectomy for uterine pathology were enrolled into the study. None of the women had ultrasound evidence of polycystic ovaries or any other ovarian pathology. One patient postponed her operation after all ovarian reserve tests had been performed, so she was excluded from the study. One other patient (patient no. 0 in Table I) was excluded because her basal level was in the post-menopausal range. All other women had a basal < ; thus patients comprised our study group. Blood samples for basal, luteinizing hormone () and European Society of Human Reproduction and Embryology Downloaded from at Pennsylvania State University on March, 0 Obstetrics and Gynaecology, Reproductive Endocrinology and Pathology, Dokuz Eylul University School of Medicine, Izmir, Turkey fecundity rates and response to the exogenous stimulation led to evaluation of other parameters. Basal serum follicle stimulating hormone () concentra tions (Scott etal, ; Toner etal, ; Farhi etal., ), clomiphene citrate challenge tests (CCCT) (Navot etal., ; Loumaye et al., 0; Scott et al., ), gonadotrophinreleasing hormone agonist stimulation tests (GAST) (Winslow et al., ), exogenous follicle stimulating hormone ovarian reserve test (EFORT) (Fanchin et al., ), elevated day serum progesterone levels (Hofmann et al., ) and inhibin B (Hall et al., ) have all been performed in order to assess response to gonadotrophins and chances of conception. Unfortunately, all the above tests are indirect measurements of ovarian reserve. Previous studies have assessed the role of these tests mainly in predicting ovarian follicular or oocyte response to gonadotrophins or chances of conception. To the best of our knowledge, the accuracy of these tests as a measure of the number of remaining primordial follicles within the ovaries has not been previously validated.

2 Accuracy of ovarian reserve tests Table I. Hormone concentrations before and after the tests and total number of the follicles and follicles per cubic centimetre Patient no. a b ± 0. Day Day Day Number of follicles Total /cc ± ± ± ± ± ± ± ± ± ± ± 6 Data omitted from further calculations because operation postponed or concentration in menopausal range respectively. oestradiol were taken from the antecubital vein on the second day of the menstrual cycle. After this procedure, mg of buserelin acetate was injected s.c. for the purpose of GAST. Twenty-four hours after the buserelin injection blood samples for,, and oestradiol were taken. After waiting for washout of buserelin for days clomiphene citrate (0 mg orally) was administered on days - and serum was obtained again on day for,, and oestradiol. Following centrifugation, sera were stored at -0 C until hormonal assessment. The operations were performed in the subsequent cycle after all tests had been completed. All three tests were performed in the same cycle but we were confident that there was no carry-over effect, especially between gonadotrophin releasing hormone (GnRH) agonist and clomiphene. Single s.c. administration of GnRH agonist (buserelin) and its half life in normal women has been studied (Lemay et al., ). It was found that (after a mg single shot of buserelin) at h, serum concentrations of both and returned to pre-treatment control values. Serum oestradiol concentrations rose progressively and reached their maximum (.-fold) at h, then began to decrease steadily, but remained above pre-treatment control values (0.-fold) at h. In the present study there was an interval of at least days between the two tests (GAST and CCCT). and were measured by a double antibody assay (coat-a count IRMA; DPC Diagnostics, Los Angeles, CA, USA). The intraand interassay coefficients of variation (CV) for the assay (for «mlu /ml) were. and.% respectively. The intra- and interassay CV for the assay (for «. mlu /ml) were and.% respectively. -ß oestradiol concentration was measured with commercially available radioimmunoassay kits (coat-a count; DPC Diagnostics). The intra- and interassay CV for oestradiol assay (for «0 pg /ml level) were and.% respectively. Basal serum concentration was considered abnormal if it exceeded. This threshold was based on the upper limit of normal (th centile) in our laboratory. Serum > on day was considered abnormal for CCCT (Tanbo et al., ). GnRH agonist stimulation test was considered normal for ovarian reserve if the serum oestradiol value on day was double that of the baseline value (on day ). Of the patients, 0 had operations for uterine myomata, one patient had uterine prolapse with urinary stress incontinence and the other had a stage cervical carcinoma. Twenty of the patients had total abdominal hysterectomy (TAH) and two patients had myomectomy. In all women > years, at the time of the study it was routine to offer oophorectomy as prophylaxis for ovarian carcinoma. In the younger women, oophorectomy was offered as treatment for premenstrual syndrome or pelvic pain. All the specimens were evaluated by the same pathologist (K.Y.). To assess the total number of follicles within the ovaries, all the ovaries were fixed for h with buffered formaldehyde and cut into slices every mm with a fractionater. All the tissues were processed in a routine manner and paraffin-embedded. From each mm slice, a mm section was cut and stained with haematoxylin-eosin (Gundersen et al., ). The sections were evaluated by light microscopy. The volume of each ovary (V( )) was calculated by the Cavalieri method (Gundersen et al., ), by multiplying the parameters from parallel sections separated by a known distance t ( mm), area associated with one point in the grid (a)(mm ), and sum of the number of points hitting the section of the ovary (S P). ov V(oV) = (t)x(a)x(s P) The sections were then evaluated for the number of follicles (primordial, primary, secondary, and Graffian follicles). Atretic follicles were not counted. The total follicle number (N( )) and ov Downloaded from at Pennsylvania State University on March, Mean ± SEM Age (years)

3 Table II. Hormonal values and number of the follicles per cubic centimetre according to the tests Test Normal Abnormal GAST n = n = Day () 6. ±.. ± 0. Day (). ±.. ±. Day oestradiol (). ±.. ±. Day (). ±.. ±.t Day (). ±.. ±. Day oestradiol () 0. ±. 6. ±. No.of the follicles/cc (mean ± SEM) ± 0 a 6 ± a CCCT n=6 n=6 Day () 6.0 ± 0.6. ±. Day (). ± 0.. ±. Day oestradiol (). ±.. ±. Day () 6. ± ±. Day () 6.0 ± 0.. ±. Day oestradiol (). ±.0. ±. No.of the follicles/cc (mean ± SEM) 0 ± b ± 6 b Basal n = 6 n = 6 Day (). ± 0.. ±.6 No.of the follicles/cc (mean ± SEM) ± c 6 ± 6 c GAST = gonadotrophin-releasing hormone agonist stimulation test; CCCT = clomiphene citrate challenge test; = follicle stimulating hormone. Differences between normal and abnormal groups: a P = 0.0; b, c P < 0.000; c P < follicle number per cubic centimetre (N( ov )/cc) of the tissue were calculated (Gundersen etal., ) using the formulae: N(ov) = (So/S v)x(v(ov)) N(ov)/cC = N(ov)/(V(ov)) where So, Sv represented counted total follicle number and calculated total section volume. At the beginning of the calculations, after the first counts, variance of S area, total variance of SP, Nugget effect (i.e. independent variance of each estimate), and Nug % was kept at.66 for point counting. The total variance of SP representing the error ratio was.%. Uni- or bilateral ovarian follicle numbers and ovarian volume as well as oocyte numbers per centimetre were calculated and the number of follicles per unit tissue was determined. Statistical Package for Social Sciences (SPSS, Corp., Chicago, IL, USA) was used for statistical analyses. The paired and unpaired t-test and simple linear regression analyses were used where appropriate. The cut-off value for the number of follicles per unit tissue was taken as the mean minus SEM. The sensitivity and specificity of each test were calculated and receiver operator curve (ROC) analysis performed using this cut-off value. Values are given as mean ± SEM. Results A total of patients with a mean age of. ± 0. (range 6 - ) years comprised our study group. The results of all the tests for each patient are presented in Table I. Table II shows the summary of results for each test. Ovarian reserve was considered as abnormal in a varying number of patients depending on the type of test performed. However, if the cut-off value for number of follicles was taken as 0 000/cc based on the mean - SEM, the ovarian reserve was abnormal in patients. For this value, the sensitivity was.% for CCCT whereas the specificity was 0.%. Sensitivity and specificity for basal were 6. and.%, and for GAST were. and.% respectively. The receiver operator characteristic curves and area under the curve indicated that the CCCT was the most accurate of the three tests assessed (Table III, Figure ). Table III. Area under the curve for each test according to ROC analysis CCCT (n = ) Basal (n = ) GAST (n = ) w (area under curve) SE w The number of follicles per unit tissue showed a negative correlation with basal (r = -0.) and with after CCCT (r = -0.) while there was a positive correlation between oestrogen after GAST (r = 0.) and follicles per unit tissue. However, none of the correlations reached statistical significance (Figure ). Also a negative correlation was noticed between age and follicle per unit tissue (r = 0., data not shown). Discussion The clinical use of gonadotrophins to induce follicular develop ment is associated with a wide variation in ovarian response. Indeed, this ovarian responsiveness to stimulation plays a major role in assisted reproductive technologies, which are complex and costly. On the other hand, poor response to ovarian stimulation may alert the clinicians to a reduction in the women's fecundity. In most cases decreased ovarian function seems to be progressive. Another difficult area for clinicians is to provide infertile couples with accurate informa tion about their chances of pregnancy. Although several factors influence the outcome of ovarian stimulation, ovarian response to stimulation is a major one in relation to IVF success. Each factor that influences ovarian response has its limitations, so ovarian reserve tests have been designed accordingly. Nevertheless, none of the screening tests that are currently Downloaded from at Pennsylvania State University on March, 0

4 Accuracy of ovarian reserve tests Figure. The receiver operator characteristic curves for basal follicle stimulating hormone (), gonadotrophin-releasing hormone agonist stimulation test (GAST), and clomiphene citrate challenge test (CCCT). available gives an appropriate answer to which test is the best for predicting diminished ovarian reserve. Traditionally, as a woman's age increases, her oestradiol response to ovulation induction and number of oocytes retrieved decreases. In fact her functional ovarian reserve plays the major role in this impaired response. This decrease in ovarian function has been attributed to an absolute reduction in the number of the follicles available for stimulation. To the best our knowledge this is the first study that directly compares ovarian reserve tests to the number of the follicles within the ovaries. The chronological age of a woman is related to her chance of conceiving and the decline in fertility after the age of years has been well documented, even in the assisted reproductive technologies (Tan et al., ). For this reason our study population was restricted to women over years of age. We are aware that a screening test should have a high sensitivity and low specificity in order not to underestimate the number of affected subjects. On the other hand, a test could only be diagnostic for a certain disease with a high specificity, although this would increase the cut-off value and allow individuals without disease to be included. It is therefore Figure. Correlation between follicles/cc with (a) GAST, (b) Basal, (c) CCCT. essential to perform receiver operator characteristic curves when comparing the accuracy of screening tests. In the present study the basal concentrations were abnormal in % (six out of ) of our patients and did not accurately reflect the actual number of the follicles per unit tissue. The sensitivity of basal was 6.% and specificity.%. In other words basal has poor predictive value to estimate the follicles per unit tissue. However, the number of our patients who had basal values < (i.e. normal) was greater than the number classified as abnormal, and this may have had an adverse effect on our results. Nevertheless, of the tests performed, basal serum was the second best in providing accurate prognostic information according to ROC analysis (Figure ). Downloaded from at Pennsylvania State University on March, 0

5 The CCCT was first used by Navot et al. to assess the ovarian reserve (Navot et al., ) and this test appears to be more sensitive than basal (Loumaye et al., 0; Tanbo et al., ; Loumaye, ). In our study, CCCT reflected a normal ovarian reserve in % (6 out of ) of our patients. The mean number of follicles per unit tissue was significantly higher in patients with normal CCCT compared to those with an abnormal result. According to the ROC analysis, the CCCT was the most sensitive (area under the Acknowledgements The authors would like to thank Mr Alp Ergor of the Public Health Department at the Dokuz Eylul University, Izmir, Turkey for his assistance with the statistics. We sincerely appreciate Professor Roy Homburg of the Tel Aviv University, Israel for critical review of the manuscript and correcting the English text. References Chang, M.Y., Chiang, C.H., Hsieh, T.T. et al. () Use of antral follicle count to predict the outcome of assisted reproductive technologies. Fertil. Steril., 6, 0-. Faddy, M.J. and Gosden, R.G. (6) A model confirming the decline in the follicle numbers to the age of menopause in women. Hum. Reprod.,, -6. Fanchin, R., de Ziegler, D., Olivennes, F. et al. () Exogenous follicle stimulating hormone ovarian reserve test (EFORT): a simple and reliable screening test for detecting 'poor responders' in in-vitro fertilization. Hum. Reprod.,, Received on March, ; accepted on August 6, Downloaded from at Pennsylvania State University on March, 0 curve = 0., Table III). The prognostic value of GAST has been documented by many authors in different protocols (Muasher et al., ; Padilla et al., 0; Winslow et al., ). Eight of our patients (6%) had normal ovarian reserve according to GAST. However, there was no statistically significant difference in the mean number of follicles per unit tissue between the groups with normal and abnormal GAST. Furthermore, GAST was the least sensitive test according to ROC analysis (area under the curve = 0., Table III), and was relatively less accurate than the other two tests. In addition, we believe that the GAST is more invasive, expensive, and needs further standardization. Some studies (Lass et al., ; Tomas et al., ; Chang et al., ) have assessed the use of ultrasound in evaluating ovarian reserve. Although ultrasound was performed in this study, ultrasonographic parameters were not an intended out come measure. We are aware of the limitations of the present study: The difficulty in recruiting patients who need an oophorectomy after years of age and before they reach 0 years played a major role in defining the number of the study population. Secondly, since all of our patients were of proven fertility, the other possible factors that can affect the ovarian reserve in an infertile population cannot be eliminated in our study group. Finally, the number of the abnormal tests was small because of the limited number of the whole study group. In conclusion, it seems unlikely that the sensitivity of the above tests can judge the ovarian reserve accurately. Nevertheless, according to ROC analysis, CCCT is more predictive of ovarian reserve compared with basal serum and GAST. Farhi, J., Homburg, R., Ferber, A. et al. () Non-response to ovarian stimulation in normogonadotrophic, normogonadal women: a clinical sign of impending onset of ovarian failure pre-empting the rise in basal follicle stimulating hormone levels. Hum. Reprod.,, -. Gundersen, H.J., Bendtsen, T.F., Korbo, L. et al. () Some new, simple and efficient stereological methods and their use in pathological research and diagnosis. APMIS, 6, -. Hall, J.E., Welt, C.K. and Cramer, D.W. () Inhibin A and inhibin B reflect ovarian function in assisted reproduction but are less useful at predicting outcome. Hum. Reprod.,, -. Hofmann, G.E., Thie, J., Scott, R.T. etal. () Evaluation of the reproductive performance of women with elevated day progesterone levels during ovarian reserve screening. Fertil. Steril.,, -. Lass, A., Silye, R., Abrams, D.-C. et al. () Follicular density in ovarian biopsy of infertile women: a novel method to assess ovarian reserve Hum. Reprod.,, -. Lemay, A., Metha, A.E., Tolis, G. et al. () Gonadotropins and estradiol responses to single intranasal or subcutaneous administration o f a luteinizing hormone-releasing hormone agonist in the early follicular phase. Fertil. Steril.,, Loumaye, E. () Ovarian challenge tests. In Hedon, B., Bringer, J., and Mares, P. (eds). Fertility and Sterility: A Current Overview. The Parthenon Publishing Group, New York, London, pp. -. Loumaye, E., Psialti, I., Billion, J.M. et al. (0) Predicition of individual response to controlled ovarian hyperstimulation by means of a clomiphene citrate challenge test. Fertil. Steril.,, -0. Muasher, S.J., Ellis, L.M., Oehninger, S. etal. () The value of basal and/ or stimulated serum gonadotropin levels in prediction ofstimulated response and in vitro fertilization outcome. Fertil. Steril., 0, -0. Navot, D., Rosenwaks, Z. and Margalioth, E.J. () Prognostic assessment of female fecundity Lancet, ii, 6-6. Padilla, S.L., Bayati, J. and Garcia, J.E. (0) Prognostic value of early serum estradiol response to leuprolide acetate in in vitro fertilization. Fertil. Steril.,, -. Scott, R.T., Oehninger, S., Toner, J.P. etal. () Follicle-stimulating hormone levels on cycle day are predictive of in vitro fertilization outcome. Fertil. Steril.,, 6-6. Scott, R.T., Leonardi, M.R., Hofmann, G.E. et al. () A prospective evaluation of clomiphene citrate challenge test screening of the general infertility population. Obstet. Gynecol.,, -. Tan, S.L., Royston, P., Campell, S. et al. () Cumulative conception and livebirth rates after in-vitro fertilisation. Lancet,, 0-. Tanbo, T., Norman, N., Dale, P.O. et al. () Prediction of response to controlled ovarian hyperstimulation: a comparison ofbasal and clomiphene citrate-stimulated follicle-stimulating hormone levels. Fertil. Steril.,, -. Tomas, C., Nuojua-Huttunen, S. and Martikainen, H. () Pretreatment transvaginal ultrasound examination predicts ovarian responsiveness to gonadotrophins in in-vitro fertilization. Hum. Reprod.,, 0-. Toner, J.P., Philput, C.B., Jones, G.S. et al. () Basal follicle-stimulating hormone level is a better predictor of in vitro fertilization performance than age. Fertil. Steril.,, -. Winslow, K.L., Oehninger, S.C., Toner, J.P. et al. () The gonadotropinreleasing hormone agonist stimulation test a sensitive predictor of performance in the flare-up in vitro fertilization cycle. Fertil. Steril., 6, -.