**3. The significance of human aging**

186 Current Frontiers in Cryobiology

The introduction of vitrification led to the first live birth after this technique in 1999 (Kuleshova et al., 1999). Vitrification of embryos has been shown to be reliable (Kolibianakis et al., 2009) and vitrification is now routinely applied to oocytes. The efficiency of both oocyte cryopreservation methods has been studied and shows similar trends in improvement. While more data exists for slow-freeze cryopreservation, due solely to the number of years that this method has been available, vitrification data is equally promising. In a 2006 meta-analysis, Oktay et al. demonstrated live birth rates increasing after slowfreezing from 21.6% per transfer (from 1996 to 2004) to 32.4% (from 2002 to 2004). The vitrification data showed a similar trend, with a live birth rate of 29.4% before 2005 and 39% after 2005 (Oktay et al., 2006). Through the use of intracytoplasmic sperm injection (ICSI), many of the concerns about zona pellucida hardening from cryopreservation were bypassed. This ability to augment fertilization of thawed oocytes altered the outlook on

oocyte cryopreservation and made this a more viable option for fertility preservation.

More recently, experimentation with cryopreservation of ovarian tissue through orthotopic or heterotopic transplantation has been attempted. The first child born after ovarian tissue cryopreservation was documented in 2004, to a woman who had a history of chemotherapy and radiation treatment for lymphoma (Donnez et al., 2004). Due to fewer studies on optimal cryopreservation protocols and methods of tissue selection, this technique has been considered more experimental than oocyte cryopreservation. In addition, optimal strategies for enhancing graft revascularization are limited in the literature and ideal tissue size has not been established. In a case series of 13 patients who underwent ovarian tissue cryopreservation due to various diseases requiring chemotherapy, large strips (8-10mm x 5mm) and small cubes (2mm x 2mm) of ovarian tissue were both effective in restoring ovarian function (Donnez et al., 2011). However, due to small numbers of human patients having undergone this procedure and a lack of standardized protocols, these outcomes are difficult to interpret. Furthermore, appropriate candidate selection for ovarian tissue cryopreservation has note been defined. Due to waning primordial follicle counts as women age, it has been suggested that ovarian tissue cryopreservation should be limited, at the very least, to women <40 years of age (Oktay, 2002). Some concern also exists about the risk of ovarian metastasis and the reintroduction of malignant cells upon transplantation of thawed

*Courtesy of Reproductive Medicine Associates of New York, LLP.*

Fig. 1. Mature human oocyte.

Human aging has been well-studied and is a known contributor to the decline in fertility experienced by women. Female fecundity, or the ability to produce offspring, declines with advancing age. This is partially due to decreased numbers of oogonia, which have a steady rate of atresia from birth, with a more rapid decline around the age of 37.5 years. Numbers of oogonia, or primordial fetal oocytes, are maximal at 20 weeks' gestation, totaling between six and seven million. At birth, this number has already declined to one to two million; a mere 400,000 oocytes remain at the beginning of puberty. While this number still seems rather high, only around 500 of these oocytes are destined for maturation and ovulation. The remainder will be lost through a highly controlled system of follicular atresia and apoptosis (Williams Gynecology, 2008 Ed.), until around 1000 oocytes remain at the time of menopause (Figure 2). Since women now live longer, a larger portion of their lives are spent in reproductive senescence, and the need for reproductive assistance due to challenges associated with diminished ovarian reserve has increased (Faddy et al., 1992).

Recent research has suggested that there may be a population of oogonial stem cells, similar to that seen in males for lifelong spermatocyte production. Several studies have pointed toward the presence of mitotically-active germline stem cells in the mammalian ovary (Johnson et al., 2004; Pacchiarotti et al., 2010; Parte et al., 2011; Zou et al., 2009). Many groups have conducted experiments which have isolated stem cells capable of sustaining oocyte and follicle production *in vitro*. While these results are controversial and disputed by some (Byskov et al., 2005), the potential for regeneration of oocytes and follicular development throughout the female life span is an exciting and promising future area in assisted reproduction.

Mathematical models have been developed in order to generate prediction rules for numbers of remaining oocytes and reproductive capacity. Oocyte atresia appears to follow a biexponential pattern, with a more rapid decline in oocyte number occurring after a critical number of 25,000 follicles remain around the age of 37.5 years (Faddy et al., 1992). According to this model, around 1000 follicles remain at the age of 51, which corresponds to the median age of menopause in the general population. Other authors have studied histological samples to identify the rate of recruitment of non-growing follicles (NGF) in human ovaries from

Oocyte Cryopreservation for the Elective Preservation of Reproductive Potential 189

respectively. Additionally, women with a BAFC of less than 4 are 37 times more likely to have

A newer marker for predicting ovarian reserve is anti-Müllerian hormone, or AMH, which has been in the literature since the early 2000s (Gruijters et al., 2003). Serum AMH levels are constant throughout the menstrual cycle, unlike FSH or E2, and are not affected by other hormone levels. Because of these relatively constant levels, AMH may be useful for predicting ovarian response to stimulation cycles for IVF; its predictive power seems to be similar to that of the BAFC (La Marca et al., 2009). Additionally, AMH is secreted in primary, preantral, and small antral follicles, which are thought to comprise the pool of ovarian reserve (Figure 3). This endocrine marker is secreted by granulosa cells and reflects the transition of resting primordial follicles to growing follicles (Sowers et al., 2008). Additionally, AMH levels diminish as an FSH-dependent dominant follicle begins to develop (Broekmans et al., 2008), reinforcing its role as a marker of preantral and small antral follicles in the pool of ovarian reserve. AMH is not, on the other hand, expressed in atretic follicles. Therefore, its levels are directly correlated to the number of viable, growing follicles that remain in the ovary. Levels of AMH decline in a predictable fashion as women near the menopausal transition, which has been studied in concordance with declining levels of inhibin-B and increasing levels of FSH (thus reinforcing the soundness of this marker as a predictor of declining ovarian reserve) (Sowers et al., 2008). There is a statistical association between AMH and FSH levels in assessing ovarian reserve. Singer et al. compared the correlation between these two hormones and found that serum AMH level is highly predictive of baseline FSH level. Using these two serum marker levels in combination

may prove to be a useful predictor of ovarian reserve (Singer et al., 2009).

Controls PCOS FHA POF

Fig. 3. Mean AMH plasma levels in patients and controls. Women with PCOS have significantly higher levels of plasma AMH and women with POF have significantly lower levels, when compared to controls and women with FHA (p<0.05). PCOS: Polycystic Ovarian Syndrome; FHA: Familial Hypothalamic Amenorrhea; POF: Premature Ovarian

their cycle cancelled (Gibreel et al., 2009).

Failure. *Adapted from Broekmans et al., 2008*.

**AMH (ng/mL)**

prenatal samples through menopause (Wallace & Kelsey, 2010). This model suggests that up to 81% of the variance in non-growing follicles is due to age alone. Interestingly, the authors' mathematical model demonstrates an increased rate of non-growing follicle recruitment until the age of 14 years old, after which NGF recruitment decreases until the menopause. Using this best-fitting asymmetric peak mathematical model, it may be possible to predict ovarian reserve in women based on age and guide discussions of fertility preservation in women seeking information about oocyte cryopreservation.

Fig. 2. Number of germ cells across the human female lifespan. Germ cells peak around 6 months post-conceptional age at a level of 6-7 million. At birth, this number has declined to around 2 million germ cells remaining in the infant ovary. Further decline occurs during the rest of the lifespan, with ~500,000 remaining at puberty and only ~1000 oocytes left at menopause.

Though we have some clinical tools to help predict a woman's reproductive capacity, including hormonal tests and the basal antral follicle count, the ramifications of human aging on reproduction are still variable and difficult to predict. Traditionally, elevated levels of basal follicle stimulating hormone (FSH) and abnormal estradiol (E2) levels have been used to guide physicians who are assessing ovarian reserve. FSH is measured in the early follicular phase of the menstrual cycle, when luteal inhibin levels decrease. Classically, it is measured on day 3 after the onset of menses. Studies have shown that a day 3 FSH level above 15 mIU/mL predicts significantly lower rates of pregnancy (Scott, 1995). Concomitant measure of E2 levels may decrease the rate of false negatives when FSH values are used alone. Estradiol should be thus be measured concurrently with day 3 FSH testing. The basal antral follicle count (BAFC) has also been used widely in the field of reproductive endocrinology and infertility to help predict ovarian reserve. BAFC <4 has a specificity of 98.7% when predicting non-pregnancy following IVF (Gibreel et al., 2009). BAFC may therefore be an appropriate measure of ovarian reserve in women undergoing infertility evaluation. Meta-analysis has also shown that BAFC of less than 4 has a sensitivity and specificity to predict cycle cancellation of 66.7% and 94.7%,

prenatal samples through menopause (Wallace & Kelsey, 2010). This model suggests that up to 81% of the variance in non-growing follicles is due to age alone. Interestingly, the authors' mathematical model demonstrates an increased rate of non-growing follicle recruitment until the age of 14 years old, after which NGF recruitment decreases until the menopause. Using this best-fitting asymmetric peak mathematical model, it may be possible to predict ovarian reserve in women based on age and guide discussions of fertility preservation in women

Fig. 2. Number of germ cells across the human female lifespan. Germ cells peak around 6 months post-conceptional age at a level of 6-7 million. At birth, this number has declined to around 2 million germ cells remaining in the infant ovary. Further decline occurs during the rest of the lifespan, with ~500,000 remaining at puberty and only ~1000 oocytes left at

Though we have some clinical tools to help predict a woman's reproductive capacity, including hormonal tests and the basal antral follicle count, the ramifications of human aging on reproduction are still variable and difficult to predict. Traditionally, elevated levels of basal follicle stimulating hormone (FSH) and abnormal estradiol (E2) levels have been used to guide physicians who are assessing ovarian reserve. FSH is measured in the early follicular phase of the menstrual cycle, when luteal inhibin levels decrease. Classically, it is measured on day 3 after the onset of menses. Studies have shown that a day 3 FSH level above 15 mIU/mL predicts significantly lower rates of pregnancy (Scott, 1995). Concomitant measure of E2 levels may decrease the rate of false negatives when FSH values are used alone. Estradiol should be thus be measured concurrently with day 3 FSH testing. The basal antral follicle count (BAFC) has also been used widely in the field of reproductive endocrinology and infertility to help predict ovarian reserve. BAFC <4 has a specificity of 98.7% when predicting non-pregnancy following IVF (Gibreel et al., 2009). BAFC may therefore be an appropriate measure of ovarian reserve in women undergoing infertility evaluation. Meta-analysis has also shown that BAFC of less than 4 has a sensitivity and specificity to predict cycle cancellation of 66.7% and 94.7%,

seeking information about oocyte cryopreservation.

menopause.

respectively. Additionally, women with a BAFC of less than 4 are 37 times more likely to have their cycle cancelled (Gibreel et al., 2009).

A newer marker for predicting ovarian reserve is anti-Müllerian hormone, or AMH, which has been in the literature since the early 2000s (Gruijters et al., 2003). Serum AMH levels are constant throughout the menstrual cycle, unlike FSH or E2, and are not affected by other hormone levels. Because of these relatively constant levels, AMH may be useful for predicting ovarian response to stimulation cycles for IVF; its predictive power seems to be similar to that of the BAFC (La Marca et al., 2009). Additionally, AMH is secreted in primary, preantral, and small antral follicles, which are thought to comprise the pool of ovarian reserve (Figure 3). This endocrine marker is secreted by granulosa cells and reflects the transition of resting primordial follicles to growing follicles (Sowers et al., 2008). Additionally, AMH levels diminish as an FSH-dependent dominant follicle begins to develop (Broekmans et al., 2008), reinforcing its role as a marker of preantral and small antral follicles in the pool of ovarian reserve. AMH is not, on the other hand, expressed in atretic follicles. Therefore, its levels are directly correlated to the number of viable, growing follicles that remain in the ovary. Levels of AMH decline in a predictable fashion as women near the menopausal transition, which has been studied in concordance with declining levels of inhibin-B and increasing levels of FSH (thus reinforcing the soundness of this marker as a predictor of declining ovarian reserve) (Sowers et al., 2008). There is a statistical association between AMH and FSH levels in assessing ovarian reserve. Singer et al. compared the correlation between these two hormones and found that serum AMH level is highly predictive of baseline FSH level. Using these two serum marker levels in combination may prove to be a useful predictor of ovarian reserve (Singer et al., 2009).

Fig. 3. Mean AMH plasma levels in patients and controls. Women with PCOS have significantly higher levels of plasma AMH and women with POF have significantly lower levels, when compared to controls and women with FHA (p<0.05). PCOS: Polycystic Ovarian Syndrome; FHA: Familial Hypothalamic Amenorrhea; POF: Premature Ovarian Failure. *Adapted from Broekmans et al., 2008*.

Oocyte Cryopreservation for the Elective Preservation of Reproductive Potential 191

opportunities and concern for pregnancy complications are all significantly higher than in men. These concerns lead to choosing between career training and childbearing, thus risking subfertility by delaying reproduction for the sake of a woman's profession. Studies at our center have evaluated motivations for and trends in elective preservation of fertility in women seeking care at a New York City infertility clinic. Women seeking elective egg freezing were likely to have a high level of education, with all women having at least a bachelor's degree and 75% holding a master's or professional degree. These women were all single, nulliparous, and the majority expressed a desire to be sure they had taken advantage of all reproductive opportunities (Gold et al., 2006). Half of women interviewed described being pressured by their "biological clock" and many wanted to freeze eggs as an "insurance policy," though did not anticipate needing to use them. Interestingly, the mean patient age was 39 years old and 65% of women had reported only recently learning about egg freezing technology. In a multicenter analysis, more than 3000 women called to inquire about fertility preservation. Of these women, those who actually completed a cycle had a significantly higher average age of 37.1 years; patients who were older than 35 had fewer cycles that resulted in the recommended number of metaphase II oocytes for cryopreservation (Frank Sage et al., 2008). This may suggest an inadequate awareness of the age-related decline in fertility that occurs as part of normal human aging. Most studies on reproductive outcomes after oocyte cryopreservation (including oocyte survival rates, fertilization rates, and number of pregnancies) have analyzed women under the age of 35 (Jain & Paulson, 2006). Because of this limitation in the body of literature on oocyte cryopreservation, providers should ideally cryopreserve oocytes in women <35 years of age. As oocyte cryopreservation becomes more publicized and accurate information about declining female fertility is disseminated, the mean age of cryopreservation may decrease. Trends in the local and national economy have been studied in relation to elective medical procedures, including oocyte cryopreservation. Costs of oocyte and embryo cryopreservation have been evaluated through the LIVESTRONG database of 154 participating reproductive centers. For the average patient, the cost of oocyte cryopreservation is around \$7,800, compared to an average of \$9,300 for embryo cryopreservation (Beck et al., 2010). The costs of fertility preservation are variable based on geography and center. In a New York City private IVF program, annual per capita income showed significant positive correlation with new consults for oocyte cryopreservation. Additionally, as annual unemployment rates increased,

the number of new consults significantly decreased (Flisser et al., 2009).

Oocyte cryopreservation has many social and ethical advantages over embryo cryopreservation. Embryo cryopreservation remains the standard recommendation for fertility preservation according to ASRM guidelines, mainly due to the amount of literature studying this technique. Single women, however, may encounter social issues with freezing embryos. The option to extend fertility without the need for a male partner or sperm donor is frequently appealing to women who are not in a long-term relationship. The discomfort of anonymity associated with sperm donors is eliminated with egg freezing. Other potential issues include decisions regarding paternity and legal obligations for patients who undergo directed sperm donation, strategies for disposing of embryos if a woman gets married later in life, and how to handle the disposition of embryos if the egg donor dies and does not have explicit advanced directives in place (Jain & Paulson, 2006). These dilemmas are all circumvented with oocyte cryopreservation. Additionally, infertility centers avoid the often difficult task of synchronizing cycles between oocyte donors and recipients, in the case of

In our NYC-based infertility clinic, women presenting for new oocyte cryopreservation consultations were retrospectively evaluated. Of the 519 women presenting for new patient consultation, approximately 1/3 initiated oocyte cryopreservation cycles. The best predictors of successful oocyte cryopreservation cycles were (in order) BAFC, day 3 FSH, and age (all p<0.05) (Barritt et al., 2010). Importantly, providers must remember that all of these tests and models attempt to predict the *quantity* of oocytes available for future reproduction. Unfortunately, tests to predict oocyte *quality* are still lacking. Models incorporating multiple variables may end up being the best predictor of ovarian reserve and ART cycle success, though many still consider age the best predictor of ovarian reserve and reproductive potential.

The risk of aneuploidy is increased in older oocytes, which leads to higher rates of chromosomally abnormal fetuses and spontaneous abortion. Approximately 15-20% of pregnancies end in spontaneous abortion, or miscarriage (Barron, 1968). Maternal age has long been recognized as a risk factor for pregnancy loss. Risk of chromosomal abnormalities, decreased fecundity, and prevalence of comorbid medical illnesses rise with increasing age – all of which may lead to spontaneous abortion (Barron, 1968). Aneuploidy is thought to affect around 20% of human oocytes (Jones, 2008). Some hypothesize that rates of aneuploidy increase with age through a "two-hit" pathway: nondisjunction followed by an inability of the oocyte to detect the chromosomal abnormality. Nondisjunction, or inappropriate chromosomal separation during meiosis I, is a leading cause of aneuploidy and increases with maternal age. Oocytes from older women may have decreased cohesive bonds between chromosomes, further predisposing them to meiotic errors (Jones, 2008). Additionally, as oocytes age, they may be unable to detect errors in recombination and sister chromatid separation.

It has been well-documented that infertility rates increase with age and that reproductive aging is primarily related to oocyte age. One prospective study demonstrated infertility rates increasing from 8% in women aged 19-26 years to 13-14% in women aged 27-34 years, and ultimately to 18% for women aged 35-39 years (Dunson, 2004). Similarly, there is a decline in success rates of fresh-cycle, non-donor oocyte IVF as a woman ages. Live birth rates per embryo transfer have been documented around 47.5% for women <35 years old, with a progressive decline to 17.0% in women 41-42 years of age, according to 2009 data from the Society of Assisted Reproductive Technologies (SART) (SART, 2009). In light of this data, strategies to preserve fertility for young women are paramount.
