Clinical Management of In Vitro Fertilization With Preimplantation Genetic Diagnosis

Ilan Tur-Kaspa, M.D.


Semin Reprod Med. 2012;30(4):309-322. 

In This Article

How to Maximize IVF-PGD Outcome

A woman's chances of having a pregnancy and a live birth by using ART are influenced by many factors, some of which are patient related and outside a fertility clinic's control such as the woman's age, infertility diagnosis, history of previous births, previous miscarriages, and previous failed ART cycles.[64] When PGD is added to IVF, additional factors may affect treatment outcome. PGD decreases the numbers of embryos available for transfer by 25 to 81%. Twenty five percent of the embryos are expected to be affected with recessive single-gene disorders, 50% with dominant mutations, 30 to 70% (depending on women's age) will be aneuploid, 75% will demonstrate unbalanced translocation, and 81% will not be suitable for ET when PGD is performed for HLA matching together with recessive mutation. Moreover, the genetic status of the woman may affect her response to ovarian stimulation. The additional high cost and the technical complexities associated with PGD should also be considered when deciding on the number of embryos to be transferred after PGD.[46,48,65,66]

Infertility Diagnosis May Affect IVF-PGD Results

PGD was originally offered to fertile couples who delivered a child with a genetic disorder.[15,16] A recent study of the cumulative reproductive outcome of 1498 couples who underwent PGD found that fertility status and parity had no significant effect on PGD outcome.[49] However, in 10 to 45% of the cases, one may find in the evaluation prior to the IVF procedure that these couples have other conditions that cause infertility and may affect the IVF outcome.[23,67] For example, severe male factor, endometriosis, hydrosalpinx, and low ovarian reserve may be diagnosed. Embryo development after ICSI performed to avoid DNA contamination when molecular genetics is planned may be different when ISCI is performed because of severe male factor. Aneuploidy rate in embryos may also depend on the sperm quality.[68] Furthermore, the couples may have social habits such as smoking, drinking, and drug use that can also affect the ART results.

Thus, couples who come for IVF because of PGD should be routinely investigated for possible infertility factors independent of their genetic diagnosis. Evaluation of the ovarian reserve should be performed by using age, antral follicle count and/or anti-Müllerian hormone, and day 3 follicle-stimulating hormone (FSH) levels to allow individual tailoring of the ovarian stimulation protocol.[46,47,65,69–71] An investigation of the uterine cavity should also be done to check for polyps, submucosal fibroids, or intrauterine adhesions. Unsuspected uterine cavity abnormalities may be found in 11 to 22% of patients.[72–74] Appropriate treatment and counseling should be performed.

Ovarian Stimulation

Is There an Optimal Number of Oocytes to Start ART? Any benefit of PGD can be demonstrated only when enough oocytes/embryos are available for biopsy.[75] ART success, even without PGD, was demonstrated to be strongly dependent on women's age and the number of oocytes retrieved. Sunkara et al[76] published their analysis of 400,135 IVF/ICSI cycles performed in the United Kingdom. They found a strong association between the number of oocytes retrieved and live-birth rate (LBR) adjusted for age. The number of oocytes that maximized the LBR was 15. LBR rose with an increase in numbers of oocytes retrieved up to 15, plateaued between 15 and 20 oocytes, and steadily declined beyond 20 oocytes. Others have shown that the optimal number of retrieved oocytes to conceive in their center was 13.[77] They reached this conclusion after analyzing 7422 ART cycles and calculating the pregnancy rates per ET and per started IVF cycle. This optimal number of 10 to 15 oocytes may explain why mild ovarian stimulation is associated with decreased pregnancy rates.[78]

Is There a Minimal Number of Oocytes to Start PGD? If the optimal number of oocytes to succeed in ART is ~10 to 15 and the PGD results will significantly decrease the number of embryos available for transfer by 25 to 81%, one may ask if there is a minimum number of oocytes with which to start PGD. The statistical uncertainties with the hidden variables important to the outcome of PGD and PGS are very high.[75] In 1998 Vandervorst et al[79] suggested that if fewer than six oocytes are expected to be retrieved, oocyte retrieval for IVF-PGD treatment should be canceled. Furthermore, if fewer than nine oocytes were to be retrieved, patients should be counseled on the poor prognosis for this cycle. This article has been quoted internationally for many years as a basis for canceling PGD when fewer than eight oocytes were expected/retrieved. New data challenge this cutoff practice.

To better counsel patients on IVF-PGD outcome, Tur-Kaspa et al[46,65,69] investigated PGD's efficiency to produce embryos suitable for ET and pregnancy, with low and high number of oocytes at different age groups. During a 4-year period (November 2002 to November 2006), 560 consecutive IVF-PGD cycles were performed at IHR/RGI in Chicago. A total of 251 PGD cycles were for aneuploidy screening (AS), 166 cycles for 51 single-gene disorders (SGD), 99 cycles for HLA-matched embryos with or without SGD with or without AS, and 44 for translocations with or without AS. Patients' ovarian reserves were estimated before treatment by age, antral follicle count, and day 3 FSH levels, and changes in the stimulation protocol or increases in medication dosage to try to improve oocyte yield were implemented accordingly. Data were analyzed for all cycles by the indication for PGD and by age. The availability of normal/unaffected embryos diagnosed by PGD, as well as the likelihood of ET and pregnancy, increased with the number of oocytes retrieved. Nevertheless, a low number of oocytes (fewer than seven) was still associated with a fair chance for ET and pregnancy, especially in young patients (<35 years of age). Therefore, Tur-Kaspa et al concluded that the practice of canceling of cycles when low number of oocytes/embryo are anticipated should be reconsidered.[46,65,69]

Verpoest et al,[49] from the same center as Vandervorst et al,[79] reevaluated the 1998 findings and summarized in 2009 the cumulative reproductive outcome of 1498 couples who underwent PGD. They reached a similar conclusion to the one recommended by Tur-Kaspa. The number of oocytes collected at retrieval significantly contributed to the reproductive outcome as an independent factor. At the same time, their analysis revealed there should be no cutoff of oocyte numbers below which PGD should be canceled. This new conclusion was in contrast with the previous report that Vandervorst et al[79] published in 1998. The larger number of cycles analyzed and the more comprehensive statistical analysis performed may explain this change.

A 2006 study on PGS for recurrent pregnancy loss also demonstrates that the recommendation to cancel PGD testing because of low numbers of embryos is not justified.[66] The authors investigated whether PGS should be continued in such patients when five or fewer embryos were achieved by IVF. All patients who consented for PGS and had five or fewer embryos on day 3 were given the option to cancel the PGS and have ET. When comparing the group that decided to cancel the PGS to the group that continued with it, the author demonstrated that the implantation and delivery rates were significantly higher in the group that continued with the PGS, even though fewer embryos were available for transfers. Unal et al[80] analyzed results of PGS cycles when one blastomere was biopsied from a total of 6098 day 3 embryos with at least seven blastomeres. Although ET and pregnancy were less frequent when fewer than six oocytes were aspirated, there were no significant difference in terms of implantation rates and clinical pregnancy once a patient had 6 to 10, 10 to 20, or >20 oocytes retrieved.

In summary, with optimal ovarian stimulation and when the patient wishes to conceive with her own eggs, IVF with PGD may be continued as long as we have at least one embryo for testing. Patients should be reminded that with lower oocyte yield, their chances of ET and pregnancy are lowered but not impossible. The practice of canceling PGD when fewer than eight oocytes were retrieved or fewer than four embryos are available for biopsy may be abandoned in most cases.

Ovarian Stimulation and Aneuploidy

Achieving maximally effective ovarian stimulation while avoiding ovarian hyperstimulation syndrome (OHSS) is a foremost concern in ART. COH is required to provide enough matured oocytes for insemination. The possible impact of exogenous gonadotropins with gonadotropin-stimulating hormone (GnRH) analogs on oocytes aneuploidy is crucial in determining the preferred COH protocol for IVF with PGD. The use of GnRH antagonists was suggested as a more patient-friendly protocol.[81,82] Munné et al[83] were first to suggest an association between exogenous FSH administration and the risk of human embryonic aneuploidy. This finding was supported by some animal models[84] as well as human in vitro studies[85] and some PGS studies.[86,87]

One randomized controlled trial (RCT)[87] suggested that mild COH using 150 IU of rec-FSH with GnRH antagonist protocol resulted in a significant lower embryonic aneuploidy rate compared with the conventional protocol of midluteal long agonist protocol with 225 IU of rec-FSH. Although this is a RCT (evidence level 1), one should be very cautious in the interpretation of this study. Although 111 women, age 22 to 37 years, were randomized, 40% (27 of 67) of the cycles of the antagonist group and 25% (11 of 44) of the long agonist cycles did not reach PGD. Only 57% of the embryos that were generated were actually tested by PGS. When only one blastomere was biopsied for PGS, 34% (16 of 47) of the embryos in the antagonist group and 33% (20 of 61) in the agonist group were normal. When two cells were biopsies from day 3 embryos for PGS, 39% (37 of 96) of the embryos were diagnosed as normal in the antagonist group compared with 28% (27 of 98) in the agonist group. The ongoing pregnancy rates per started cycles were 19% in the antagonist group and 17% in the agonist group. Although this RCT was well designed, because of the low numbers of embryos that were actually tested for PGS, its result must be validated by other RCTs.

However, several studies from leading centers in the United States[88] and Europe[80,89,90] could not demonstrate a significant association between COH and aneuploidy rates. They retrospectively analyzed a combined sum of >14,000 oocytes/embryos and failed to find any significant effect of the type of medication used (pure FSH versus preparations with luteinizing hormone activity), the use of GnRH agonist versus antagonist, or the numbers of oocytes retrieved on aneuploidy rate.

Tur-Kaspa et al[88] analyzed 221 consecutive PGS cycles of patients <43 years of age with 2132 oocytes/embryos tested. The average rates of aneuploidy for women <35 years were 53 to 56% whether 1 to 7, 8 to 15, 16 to 20, and >21 oocytes were retrieved compared with 68 to 72% aneuploidy rates for women 36 to 42 years, respectively (p < 0.05). The number of oocytes retrieved, type of gonadotropins or GnRH analog used, antagonist or agonist, did not affect the aneuploidy rate. Kahraman's group have summarized aneuploidy rates from a total of 6098 day 3 embryos and found similar aneuploidy rate irrespective of the number of oocytes retrieves.[80] A recent retrospective study from an experienced center in Brussels[89] specifically looked at the possible effect on aneuploidy rate of long GnRH agonist use versus antagonist protocol. The study involved 694 consecutive PGS cycle for women <37 years of age. The aneuploidy rates were 50% in both groups, whether agonist or antagonist were used. Multivariate analysis showed that the type of stimulation adjusted for age, total gonadotropins dosage, and the numbers of oocytes retrieved did not influence the aneuploidy rate. Gianaroli et al,[90] from another leading European center, investigated aneuploidy rates in 3816 first polar bodies in 706 cycles and found no significant correlation between the proportions of normal oocytes and the type of stimulation protocol used, agonist or antagonist. An inverse and significant correlation was found between the proportion of normal oocytes and the number of FSH units used for COH per oocytes.

Poor responder patients are usually treated with a maximal dose of gonadotropins to try and increase the number of oocytes retrieved. Even in this specific challenging group, the aneuploidy rate was similar to control, suggesting that women responding poorly to COH are not at higher risk of producing aneuploid embryos in vitro.[91,92] The Brussels group[93] have retrospectively investigated factors affecting the outcome of a total of 2753 PGD and PGS cycles. Although age had a significantly negative effect on outcome, as expected and similar to regular ART, mode of inheritance, fertility status, and type of ovarian stimulation protocol did not influence the PGD success rates.

If COH will increase oocyte aneuploidy, an increased aneuploidy rate is also expected in the miscarriages after COH/intrauterine insemination (IUI) and/or IVF treatments. Massie et al[94] recently answered this question by examining the rate of aneuploidy in missed abortions among infertile couples after conceiving either through COH for IUI or IVF, or by natural conceptions. The rate of abnormal results in the cytogenetic analysis of the product of conceptions in pregnancies conceived with COH (63%) was similar to the rate in the spontaneous conceptions (70%). Moreover, despite a significant higher dosage of FSH used in the IVF group compared with the IUI patients, the miscarriage aneuploidy rate was 63% in both groups. This suggests that exogenous FSH used for COH does not increase the risk of aneuploidy.[89] Even in chromosomally normal infertile individuals, there appears to be an increased propensity to meiotic errors leading to aneuploidy.[95] Munné et al[96] compared aneuploidy rates in PGD embryos from infertile patients and normal oocytes donors (ages 18 to 34 in both groups) and found a significantly higher frequency of normal embryos in the donor group (43%) compared with the infertile group (34%), suggesting that infertile women produce more aneuploid embryos. Verpoest et al[97] demonstrated a 36.4% (95% confidence interval, 10.9 to 69.2) aneuploidy rate in embryos obtained with unstimulated IVF cycles (mean women's age: 31.4 years). Thus, infertile women should be counseled that infertility by itself could increase the probability of aneuploidy.

In summary, based on considerable data obtained from retrospective studies (>14,000 oocytes/embryos) from experienced European and U.S. centers, there is insufficient evidence to suggest that the type of gonadotropins and GnRH analogs used for ovarian stimulation protocols or the number of oocytes retrieved affect embryos' aneuploidy rate. However, inadequate stimulation and/or poor PGS/PGD techniques may indeed affect cycle outcome.

Genetic Status of the Woman and Her Response to Controlled Ovarian Hyperstimulation

The genetic diagnosis of the PGD patient may sometimes affect ovarian response to stimulation and thus may decrease the numbers of oocytes retrieved. One must remember that if these mutations, such as fragile X syndrome or myotonic dystrophy (DM), had been associated with infertility, they would not have become as frequent as they are. Polar bodies-based PGD may be performed for these or other mutations.[98,99] These young women may further demonstrate that there is no direct relationship between oocyte quantity and embryo quality.[100] Whether DM patients and female carriers of balanced translocations demonstrate lower response to COH is still debatable.

Fragile X Syndrome Platteau et al[101] were the first to demonstrate that PGD for fragile Xa syndrome may be difficult but not impossible with regard to ovarian responsiveness to COH. Because women with a permutation are at increased risk of premature ovarian failure, it was no surprise that ~20 to 30% of these women responded poorly to COH. Nevertheless, once they achieved ET, their chance of conceiving is comparable with other patients. Infertility is usually not a presenting symptom of fragile X syndrome carriers, and it seems that the decrease in the numbers of oocytes retrieved represents a quantity issue rather than a quality problem with regard to implantation.

Myotonic Dystrophy DM was also suggested to decrease ovarian response to COH.[102] Oligomenorrhea, miscarriages, and early menopause have been reported in women with myotonic dystrophy type 1 (DM1).[103] Male DM1 patients are known to have reduced sperm quality as a result of gonadal atrophy, and they have an increased risk of infertility.[104] Feyereisen et al[102] compared the ovarian response of carriers of X-linked disorders to those with DM. Although patient characteristics and the number of oocytes retrieved were similar, the numbers of days on COH were significantly prolonged and the prevalence of poor quality embryos was higher in the DM group. Sahu et al[105] also demonstrated a reduced ovarian reserve and ovarian response to COH for these women.

In contrast, Verpoest et al[106,107] found no evidence of decreased ovarian response to stimulation in women with DM1. They further demonstrated that the reproductive outcome for these women undergoing PGD was not affected by the size of the expanded CTG repeats. It seems that although some of these patients may have low response to COH, it is from resistance to stimulation and thus represents a quantity issue rather then poor oocytes quality.

Female Carriers of Balanced Translocation Female carriers of balanced translocation may also have lower ovarian response to COH. Chen et al[108] demonstrated a significantly higher proportion of female carrier of balanced translocation who responded poorly to ovarian stimulation compared with women whose male partner had the translocation. They suggested that given the significant embryo attrition due to chromosomal imbalance, aggressive stimulation should be considered if the patient is not at risk for OHSS. We have looked at the clinical outcome of all consecutive oocyte retrievals that were performed at our center in a 4-year period for PGD for reciprocal translocations (Tur-Kaspa, unpublished data). When the translocation was of female origin, compared with cycles with male origins, even at age <35 years and with higher total dose of gonadotropins used (4787 ± 1646 versus 2971 ± 1280 IU), significantly fewer oocytes were retrieved (13.5 ± 3.5 versus 18.7 ± 6.6), fewer cycles resulted in ET (25% versus 77%), and fewer embryos were transferred (1.0 ± 0 versus 1.9 ± 0.5). Here again, once the patients had an ET, the clinical pregnancy rate per ET was good and reached 44%. Because PGD for translocations decreases the number of embryos available for ET by ~75%, we considered female carriers for translocation as potential low responders, and we increase the initial dosage of gonadotropins by 75–150 IU of rec-FSH, compared with other age-matched women, to try to improve oocyte yield.

However, Benner et al[109] found no differences in ovarian stimulation parameters and cycle outcomes when the translocation was of female origin compared with cycles with male origins. Their conclusion was that female carriers of balanced translocations have no diminished ovarian reserve.

In summary, the numbers of oocytes retrieved is a significant predictor of IVF-PGD cycle success, similar to regular ART cycles. Optimal oocyte yield seems to be ~15, 10 to 15 for the younger patients, and 15 to 20 for the older ones. Nevertheless, once patients were adequately stimulated, and wish to conceive with their own eggs, oocyte retrieval and PGD may be continued even with a very low number of oocytes or embryos (one to seven). When a young patient responds poorly to COH, there is no risk for OHSS, and gonadotropins dosage may be safely increased to the maximal dosage used by the center. Patients with fragile X, DM, or balanced translocation should be informed that although the ovarian response to COH may be decreased, the outcome of ETs is not affected. The possible role of mild stimulation for patients who come for IVF because of PGD remains to be investigated.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.