Most treatment decisions in female fertility care are based on the assessment of ovarian function. A diagnosis of diminished ovarian function, often also called diminished ovarian reserve, will almost uniformly result in redirection of fertility treatments. When fewer than 25,000 follicles remain, human ovaries enter a stage of functional decline characterized by further rapid loss of follicles.1 Tests for diminished ovarian reserve are used to identify women who have entered this stage and are, therefore, less likely to benefit from in vitro fertilization (IVF).2–5 Women with diminished ovarian reserve are then often advised to use donated eggs to achieve pregnancy or to adopt a child.
So far, diminished ovarian reserve has been poorly defined, with various authors using different criteria for diagnosis. Two such criteria have stood out over the years: the basal follicle-stimulating hormone (b-FSH) level, usually drawn on days two or three of the menstrual cycle, and the diagnosis of resistance to ovarian stimulation with gonadotropins. Both have been used in rather rigid fashion in that neither has been considered in the context of female age. For example, b-FSH levels of less than 10 to 12 milli international units/mL are considered normal6–8 and b-FSH between 12 and 15 milli international units/mL are often considered sufficient criteria for referral to a donor egg program.9
Because most biologic functions exist on a continuum, it is reasonable to assume that diminished ovarian reserve actually represents a range of decreasing ovarian function, from mild to severe. We have recently suggested that some women with so-called unexplained infertility may actually suffer from early diminution of ovarian function.10 Others also proposed that tests of ovarian reserve could be used to detect early ovarian aging among otherwise asymptomatic young women,11,12 suggesting that early ovarian aging affects up to 10% of women.13 Weghofer et al12 demonstrated that, even within apparently normal b-FSH levels, women with higher b-FSH levels produced reduced numbers of oocytes in IVF.
Because current criteria for the diagnosis of diminished ovarian reserve, whatever measure are used to reach the diagnosis, usually encompass all age groups indiscriminately, they most likely in younger women are too generous, and, possibly, too restrictive for older patients. As Weghofer et al12 suggested, younger patients probably experience some degree of diminished ovarian function long before their b-FSH levels reach 10–12 milli international units/ mL, while some older women, even at relatively higher b-FSH levels, may still demonstrate a better than average oocyte response for their age.
We, therefore, hypothesized that at each age there are critical values of b-FSH that can identify women with suboptimal ovarian function, characterized by poor response to standard ovarian stimulation. We choose to call this state premature ovarian aging. The objective of this study was, therefore, to characterize the continuum of ovarian function based on age-specific categories of b-FSH
MATERIALS AND METHODS
We used a convenience sample of patients with apparently normal ovarian function for their respective ages, listed in our center's Society for Assisted Reproductive Technology database between May 1999 and April 2006. To be considered to have normal ovarian function, these patients had to be classified as such in the data base, had to be aged 41 years or younger, could not have demonstrated a b-FSH level of 12 milli international units/mL or more, and had to be stimulated in their IVF cycle with a standard protocol involving down regulation with a long luteal phase agonist (Lupron, TAP Pharmaceuticals Inc. Lake Forest, IL) followed by ovarian stimulation with a modal value of 300 international units of gonadotropins per day. Baseline and IVF treatment cycle data were collected from 966 individual patients. Only the first cycle of treatment was selected for each patient.
Women with abnormally elevated baseline estradiol (E2) levels were excluded, as were subjects with b-FSH of 4 milli international units/mL or less and no recorded baseline E2, because their true b-FSH status remained ambiguous. After exclusions, 640 subjects remained.
In these patients, we performed linear regression of age against their b-FSH to test for a continuous linear relationship. The association of each of these variables with increasing patient age was tested by performing linear regression. The patients were divided into quartiles based on age. We compared b-FSH, cancellation rate, number of oocytes, total embryos transferred, and embryos cryopreserved among age quartiles. Categorical variables were cross-tabulated and evaluated using χ2 testing.
The 95% upper confidence limit of the mean (95% confidence interval [CI]) log normalized b-FSH and total oocytes retrieved within quartiles of age were determined to define cutoffs for each age group. We then categorized subjects within each age group based on their b-FSH level. Subjects with b-FSH level below the upper 95% CI of the mean for each age category were characterized as appropriate ovarian aging. Subjects with b-FSH equal to or greater than to the 95% CI of each age category (although still less than 12 milli international units/mL) of FSH were characterized as having diminished age-adjusted ovarian reserve and were considered to suffer from premature ovarian aging (Fig. 1). One hundred sixty women aged older than 41 years were excluded as age-appropriate diminished ovarian reserve, and an additional 46 women with b-FSH of 12 milli international units/mL or greater were defined as prematurely diminished ovarian reserve. Both groups were excluded from analysis, leaving 434 women, aged younger than 41 years, for further analysis. Patients were characterized as low oocyte producers as follows: younger than 33 years, 6 oocytes or fewer; aged 33 years or older and younger than 37 years, four oocytes or fewer; aged 38 years or older and younger than 41 years, three oocytes or fewer. We also identified the patients within each age group who produced four oocytes or fewer.
Differences in outcomes between high b-FSH and low b-FSH were tested using analysis of variance of total oocytes, and of total normal embryos within each age quartile. We cross-tabulated variables in age strata and used Mantel-Haenszel tests of the risk of producing “four oocytes or fewer” and age-adjusted “low oocyte” production in the presence of premature ovarian aging across the three age categories and calculated the sensitivity, specificity, and positive and negative predictive value of premature ovarian aging for “four oocytes or fewer” and age-adjusted “low oocyte” production.
Endocrine testing was performed in our clinical endocrine laboratory. From 1999 until mid 2005 all assays were performed on an ACS 180SE Chemiluminescent analyzer (Bayer Diagnostics Corp., Norwood, Massachusetts). From mid 2005 until the present all assays were performed on an AIA-600II (TOSOH Bioscience, Inc, Tokyo, Japan). The coefficient of variation of the assays, as determined by summation of semiannual quality control proficiency testing over this entire period has been 8.1% for the FSH assay and 8.3% for the E2 assay.
All statistical calculations were performed using SPSS for Windows, Standard version 15.0 (SPSS Co., Chicago, IL). Continuous variables were tested for normality using the Shapiro-Wilks W test and by graphic analysis using histograms, probability-probability and quantile-quantile plots. Variables that were noted not to be normally distributed were normalized by log conversion or by taking the square root. Tests of significance were carried out on both normalized and original data. If both analyses were significant only the original data are presented. Multivariable analysis was performed with the general linear model module of SPSS. Continuous variables are presented as mean ±1 standard error. Variables that are the result of log-normalized data are presented as mean and 95% CI of the mean. The statistical analysis was carried out by the lead author (D.B.). This was a review of an existing database and as such did not require institutional review board approval.
We initially reviewed 1,353 cycles. After limiting the analysis to only the first cycle of treatment for each patient, a study group of 966 patients remained. After limiting the remaining 640 patients to those in the lower three quartiles and further eliminations for incomplete data, the study group consisted of 434 patients.
There was a direct linear relationship between b-FSH and age (r2=0.07, P<.001; Fig. 1). The average age of the patients was 36.7 (95% CI 36.3–37.0) years. The mean b-FSH was 7.84 (95% CI 7.5–8.2). The mean count of oocytes produced was 10.1 (95% CI 9.7–10.6). Univariable comparisons among the key variables across age are shown in Table 1.
Total gonadotropin dosages, b-FSH, and the cancellation rate increased significantly with increasing age. In contrast, total oocytes retrieved and total normal embryos (cryopreserved plus transferred embryos), all decreased with increasing age (Table 1).
Four hundred thirty-four patients with appropriate ovarian aging, diminished ovarian reserve, or premature ovarian aging are listed by age category in Table 2. As the table demonstrates, a finding of premature ovarian aging declined from 49.7% of the infertile women aged younger than 33 years, to 42.4% at ages 33 to 37 years, and 27.7% at ages 38 to 41 years. Although there was a nominal increase in cancellation rates between appropriate ovarian aging and premature ovarian aging categories, it did not achieve significance. Total retrieved oocytes (P<.001) and total normal embryos (P<.05) were significantly decreased among patients with premature ovarian aging in all age groups. Moreover, a finding of premature ovarian aging in all age groups significantly increased the probability of egg retrieval with four oocytes or fewer (Fig. 2), with a positive predictive value of 19.5% and a negative predictive value of 88%. We note that the women in the 33- to 36-year-old age group with premature ovarian aging had the same risk of producing four oocytes or fewer as appropriate ovarian aging 37- to 41-year olds. Similarly, 37- to 41-year-old women with premature ovarian aging had the same risk of producing four oocytes or fewer as older women with classical diminished ovarian reserve.
The Mantel-Haenszel common odds ratio for “four oocytes or fewer” production and “low oocyte” production in the presence of premature ovarian aging were 2.8 (95% CI 1.52–5.17; P<.001) and 2.4 (95% CI 1.55–3.62; P<.001) respectively. The positive and negative predictive values, sensitivity, and specificity of premature ovarian aging for “4 oocyte or fewer” production and “low oocyte” production within each age category are shown in Table 3.
In this study, we examined the ovarian response to ovulation induction at varying ages within traditionally defined limits of normal ovarian function. Our results suggest that, within generally accepted normal b-FSH values, women with b-FSH above the 95% confidence limits for age produce fewer oocytes in response to normal ovulation induction protocols compared with other women their age. Although we did not find a significant increase in cancellation rate among women with premature ovarian aging, it may be that because women themselves participate in the decision to cancel a cycle, this criterion may have less objectivity than the other outcomes measured. In this report, we have identified 175 women with premature ovarian aging of 434 who would have been otherwise classified as “normal” by standard criteria. We have shown that these women are at a significant risk of producing fewer oocytes than their peers. Current fertility practice standards fail to take note of these women with evidence of declining ovarian function. As we previously suggested, such women are often misdiagnosed as suffering from so-called unexplained infertility.10 In Figure 2 the proportion of premature ovarian aging patients with four oocytes or fewer per retrieval in the second and third age categories approximates, or is greater than the proportion of women with 4 or fewer oocytes per retrieval in the next higher age category with age-appropriate ovarian aging b-FSH.
Identification and optimal treatment of young poor responders and older good responders is a continuing area of interest in our specialty.6,14,15 Currently b-FSH levels of greater than 12 to 15 milli international units/mL are widely accepted as laboratory definition of diminished ovarian reserve and as a predictor that women will be poor responders to ovulation induction.4,6–8,16,17 Some other definitions of “poor response” are based on fewer than three to five follicles on the day of hCG administration, fewer than three to five retrieved oocytes,15 peak E2 levels of less than 300 to 500 pg/mL,18 anti-müllerian hormone,19 or antral follicle count.20–22
The interval from onset of rapid decline of ovarian function, at approximately age 37.5 years, to menopause is estimated to be about 13 years1 and the interval from onset of menstrual irregularity to menopause is estimated to be approximately six years.23 Falling below age-expected ovarian responsiveness to ovulation induction has been associated with an earlier onset of menopause24–26 and, so far, has been considered to be the earliest sign of decline in ovarian function.11 This study suggests that such a diagnosis can be reached earlier.
The primary intent of existing pretreatment screening of ovarian reserve is to identify women who will benefit from fertility therapy by establishing pregnancy and reaching delivery. The usefulness of such screening has, however, recently been questioned. For example, Mol et al27 calculated that the sensitivity and specificity of pretreatment ovarian reserve testing methods had to improve to 50% and 96%, respectively, to be clinically valuable as a predictor of pregnancy and live birth.
Production of normal oocytes and high-quality embryos is a prerequisite step to achieving ongoing pregnancy. In the present study, the use of outcomes of normal oocyte and embryo production as endpoints provides greater statistical power to evaluate predictors of performance and alternate treatments. Although past studies may not have found a clinically significant predictive value of non–age-specific b-FSH, in the here reported study, we demonstrate that age-specific b-FSH effectively predicts response to ovulation induction (Table 3).
In recent years ovarian reserve screening has also been used to identify individuals who might benefit from more aggressive ovulation induction.28 Whether such an approach is effective in increasing oocyte yield and pregnancy rates, has remained controversial as some authors have failed to detect such benefits in traditionally defined poor responders.29,30 Others, however, have reported improvements in response with high-dose gonadotropins, combined with agonist flare protocols.31 Although at this point still speculative, such conflicting results suggest that aggressive ovarian stimulation may still benefit milder-affected women, although severely compromised ovaries may no longer be able to respond favorably. Although this subject quite obviously requires further investigation, age-specific b-FSH testing may offer a new tool to determine which patients may, or may not, benefit from stronger stimulation
In a continuum of responses to standard ovulation induction, some women will demonstrate normal age-specific response and others abnormally low response. Among the latter, only a subgroup will fit the historically used definitions of poor response. In the present study, we identify a group of women, with b-FSH below the historical threshold of 12 milli international units/mL but above the 95% CI of mean age-specific b-FSH, who produce significantly fewer oocytes than their age-comparable counterparts with appropriate ovarian aging ovarian function after standard ovarian stimulation. Based on their diminished response to standard ovulation induction, we consider these women to have evidence of incipient diminished ovarian reserve at prematurely young ages and have, therefore, designated them as suffering from premature ovarian aging.
Such a diagnosis is not only of academic interest: It often identifies young women who, otherwise, would either go undiagnosed, or wrongly diagnosed with unexplained infertility.10 In addition, a timely diagnosis of premature ovarian aging facilitates timely intervention. If such patients are treated with stimulation protocols, based on their “ovarian age,” pregnancy rates will be surprisingly high in comparison to historically reported rates in such patient populations.32 Indeed, following such an approach, we were unable to detect significant differences in initial pregnancy rates between age-matched premature ovarian aging patients and controls with age-appropriate ovarian function, although the cumulative pregnancy chance, considering fewer cryopreserved embryos, was statistically lower among premature ovarian aging patients.32
This study, therefore, suggests that chronological female age is only second best in predicting ovarian response to ovulation induction. What, indeed, seems to represent a more reliable measure of ovarian function is “ovarian age,” ie, the ability to diagnose younger women with inappropriately aging ovaries at a time when, under current practice patterns, such women would go largely unrecognized in their ovarian dysfunction. If such women then are treated with “ovarian age-appropriate” ovarian stimulation, IVF should give them outstanding pregnancy success. Because patient populations and laboratory procedures may differ, it is important for each practice to determine the appropriate cutoff that will define premature ovarian aging among their own patients.
Lobo33 has previously called for the specific targeting of women, perceived to be “at risk” for early ovarian aging. This seems particularly opportune at a time when pregnancy in most developed countries is rather uniformly delayed.34 As diagnostic testing of ovarian function improves and young women at greatest risk for premature ovarian aging can be reliably identified, they then can be given the opportunity to adjust reproductive planning accordingly. This study suggests that age-specific ovarian function testing represents considerable progress toward such a goal.
1. Faddy MJ, Gosden RG, Gougeon A, Richardson SJ, Nelson JF. Accelerated disappearance of ovarian follicles in mid-life: implications for forecasting menopause. Hum Reprod 1992;7:1342–6.
2. Levi AJ, Raynault MF, Bergh PA, Drews MR, Miller BT, Scott RT Jr. Reproductive outcome in patients with diminished ovarian reserve. Fertil Steril 2001;76:666–9.
3. Navot D, Rosenwaks Z, Margalioth EJ. Prognostic assessment of female fecundity. Lancet 1987;2:645–7.
4. Scott RT, Toner JP, Muasher SJ, Oehninger S, 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.
5. Scott RT, Opsahl MS, Leonardi MR, Neall 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.
6. Fasouliotis SJ, Simon A, Laufer N. Evaluation and treatment of low responders in assisted reproductive technology: a challenge to meet. J Assist Reprod Genet 2000;17:357–73.
7. 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.
8. Kligman I, Rosenwaks Z. Differentiating clinical profiles: predicting good responders, poor responders, and hyperresponders. Fertil Steril 2001;76:1185–90.
9. Toner JP, Flood JT. Fertility after the age of 40. Obstet Gynecol Clin North Am 1993;20:261–72.
10. Gleicher N, Barad D. Unexplained infertility: does it really exist? Hum Reprod 2006;21:1951–5.
11. Nikolaou D, Templeton A. Early ovarian ageing: a hypothesis. Detection and clinical relevance. Hum Reprod 2003;18:1137–9.
12. Weghofer A, Margreiter M, Fauster Y, Schaetz T, Brandstetter A, Boehm D, et al. Age-specific FSH levels as a tool for appropriate patient counselling in assisted reproduction. Hum Reprod 2005;20:2448–52.
13. Nikolaou D, Templeton A. Early ovarian ageing. Eur J Obstet Gynecol Reprod Biol 2004;113:126–33.
14. Tarlatzis BC, Zepiridis L, Grimbizis G, Bontis J. Clinical management of low ovarian response to stimulation for IVF: a systematic review. Hum Reprod Update 2003;9:61–76.
15. Surrey ES, Bower J, Hill DM, Ramsey J, Surrey MW. Clinical and endocrine effects of a microdose GnRH agonist flare regimen administered to poor responders who are undergoing in vitro fertilization. Fertil Steril 1998;69:419–24.
16. Cameron IT, O'Shea FC, Rolland JM, Hughes EG, de Kretser DM, Healy DL. Occult ovarian failure: a syndrome of infertility, regular menses, and elevated follicle-stimulating hormone concentrations. J Clin Endocrinol Metab 1988;67:1190–4.
17. Chuang CC, Chen CD, Chao KH, Chen SU, Ho HN, Yang YS. Age is a better predictor of pregnancy potential than basal follicle-stimulating hormone levels in women undergoing in vitro fertilization. Fertil Steril 2003;79:63–8.
18. Brzyski RG, Muasher SJ, Droesch K, Simonetti S, Jones GS, Rosenwaks Z. Follicular atresia associated with concurrent initiation of gonadotropin-releasing hormone agonist and follicle-stimulating hormone for oocyte recruitment. Fertil Steril 1988;50:917–21.
19. van Rooij IA, Broekmans FJ, te Velde ER, Fauser BC, Bancsi LF, de Jong FH, et al. Serum anti-Mullerian hormone levels: a novel measure of ovarian reserve. Hum Reprod 2002;17:3065–71.
20. Frattarelli JL, Levi AJ, Miller BT, Segars JH. A prospective assessment of the predictive value of basal antral follicles in in vitro fertilization cycles. Fertil Steril 2003;80:350–5.
21. Nahum R, Shifren JL, Chang Y, Leykin L, Isaacson K, Toth TL. Antral follicle assessment as a tool for predicting outcome in IVF–is it a better predictor than age and FSH?. J Assist Reprod Genet 2001;18:151–5.
22. Chang MY, Chiang CH, Hsieh TT, Soong YK, Hsu KH. Use of the antral follicle count to predict the outcome of assisted reproductive technologies. Fertil Steril 1998;69:505–10.
23. den Tonkelaar I, te Velde ER, Looman CW. Menstrual cycle length preceding menopause in relation to age at menopause. Maturitas 1998;29:115–23.
24. Nikolaou D, Lavery S, Turner C, Margara R, Trew G. Is there a link between an extremely poor response to ovarian hyperstimulation and early ovarian failure? Hum Reprod 2002;17:1106–11.
25. Lawson R, El-Toukhy T, Kassab A, Taylor A, Braude P, Parsons J, et al. Poor response to ovulation induction is a stronger predictor of early menopause than elevated basal FSH: a life table analysis. Hum Reprod 2003;18:527–33.
26. de Boer EJ, den Tonkelaar I, te Velde ER, Burger CW, Klip H, van Leeuwen FE, et al. A low number of retrieved oocytes at in vitro fertilization treatment is predictive of early menopause. Fertil Steril 2002;77:978–85.
27. Mol BW, Verhagen TE, Hendriks DJ, Collins JA, Coomarasamy A, Opmeer BC, et al. Value of ovarian reserve testing before IVF: a clinical decision analysis. Hum Reprod 2006;21:1816–23.
28. Hofmann GE, Sosnowski J, Khoury J, Kearns L, Thie J. Obstetric implications for and identification of women with a normal clomiphene citrate challenge test result who have a poor response to gonadotropins. Fertil Steril 2004;82:908–12.
29. Karande VC, Jones GS, Veeck LL, Muasher SJ. High-dose follicle-stimulating hormone stimulation at the onset of the menstrual cycle does not improve the in vitro fertilization outcome in low-responder patients. Fertil Steril 1990;53:486–9.
30. Yong PY, Brett S, Baird DT, Thong KJ. A prospective randomized clinical trial comparing 150 IU and 225 IU of recombinant follicle-stimulating hormone (Gonal-F*) in a fixed-dose regimen for controlled ovarian stimulation in in vitro fertilization treatment. Fertil Steril 2003;79:308–15.
31. Karande V, Morris R, Rinehart J, Miller C, Rao R, Gleicher N. Limited success using the “flare” protocol in poor responders in cycles with low basal follicle-stimulating hormone levels during in vitro fertilization. Fertil Steril 1997;67:900–3.
32. Gleicher N, Barad D. “Ovarian age-based” stimulation of young women with diminished ovarian reserve results in excellent pregnancy rates with in-vitro fertilization. Fertility and Sterility 2006;86:1621–5.
33. Lobo RA. Early ovarian ageing: a hypothesis. What is early ovarian ageing? Hum Reprod 2003;18:1762–4.
34. Heck KE, Parker JD, McKendry CJ, Chavez GF. Mind the gap: bridge methods to allocate multiple-race mothers in trend analyses of birth certificate data. Matern Child Health J 2003;7:65–70.