The International Council on Infertility Information Dissemination, Inc

Reading Between the Lines in the CDC IVF Clinic Statistics Reports by Sam Thatcher, M.D., Ph.D.


Reading Between the Lines in the CDC IVF Statistics Clinic Reports

By Sam Thatcher M.D., Ph.D. 

The Fertility Clinic Success Rate and Certification Act of 1992 requires that U.S. clinics performing assisted reproductive technology (ART), fertility treatments in which both eggs and sperm are handled, report their success rates to the Centers for Disease Control (CDC). The CDC publishes an annual report detailing the ART success rates for each of these clinics and also summarizes the data from all clinics collectively. This information is widely used and is available at


Couples considering ART should be cautious of programs that report success rates significantly lower, or even higher, than the national average. Couples should also understand that the difference in clinic specific success rates might be related more to patient selection and aggressiveness of therapy than to program quality. A direct comparison of different clinics by success rates alone may lead to erroneous conclusions. It is imperative that each couple has a frank discussion with their physician so that their individual chances of success can be placed in the framework of an individual clinic’s success.


In 2001, there were 421 ART clinics identified in the United States and 384 submitted data to the CDC regarding 107,587 cycles and about 40,000 babies. For all ages the success rates were 32.8% pregnancies per cycles and 33.4% live births per transfer. The following is an editorial explanation of how to interpret the CDC reports. To get the most benefit from this presentation, one should have a copy of a specific clinic's report as reference. Third party reproduction (donor egg, donor embryo, and gestational carrier) is not presented for brevity.


1.     Types of ART

 While the GIFT and ZIFT procedures had arguably better success rates than IVF and embryo transfer in the distant past, these types of ART are now generally restricted to very specific indications. Only about 1 % of assisted reproduction was not related to IVF. In 2001, success with GIFT was less than with IVF. Clinics performing more than the occasional GIFT/ZIFT procedure should be questioned about the rationale for their approach. Many large clinics perform no ART other than IVF or IVF/ICSI.


There were only a very small number of total cycles in which no stimulation was given, probably a few more in which Clomiphene was used. The vast majority of cycles use gonadotropin stimulation.


Virtually all ART clinics now perform intracytoplasmic sperm injection (ICSI) and it is a part of about 50% of 2001 IVF cycles, but there is great variance among clinics in its utilization. Some programs are very liberal and perform as much as 80% ICSI while others perform less than 10%. Some clinics limit ICSI to severe well-documented male factor cases while others may expand it to include unexplained infertility as well as borderline cases of male factor. While ICSI has revolutionized therapy for male infertility and allows pregnancies for many who otherwise would be relegated to use of donor sperm, it remains controversial whether ICSI should be performed without a specific reason. About 50 % of ICSI treatments had no male factor identified. Overall success rate with ICSI was slightly lower than conventional IVF with and without male factor. There is little evidence that ICSI is beneficial for issues of egg quality alone.


 2. Patient Diagnosis 
Initially, IVF was used in treatment of tubal disease where there was an obvious rationale to bypass the need for the fallopian tubes. IVF was quickly extended to treatment of male factors because of the equally obvious benefit of allowing larger number of motile sperm to be placed in closer proximity to the egg. Now, IVF represents the final common pathway for virtually all forms of infertility. In some cases, IVF may be appropriately utilized as a first line approach, while in others it may come at the end of an exhaustive evaluation and multiple interventions. IVF may correct a specific block to fertility, it may allow multiple causes of infertility to be treated at one time, or it may simply overpower the cause of infertility by its aggressiveness.


Populations of patients may vary in different clinics and areas of the country.

Individual clinics do not report their success rates by each type of infertility. The CDC tabulates all indications and reports success with different types of infertility for all clinics combined. The worst prognoses are for the individuals with diminished ovarian reserve (decreased egg supply). This may occur with advancing age and when endometriosis, surgery, or other factors have damaged the ovary. Each clinic may define diminished ovarian reserve differently. The CDC reports quite good success rates with treatments for endometriosis, but there is no separation between its milder forms, which should have excellent chances of success and severe forms in which multiple reports have shown significant compromise in success. Individuals with multiple factors do not perform as well as those with single factor infertility. The greatest success is found where IVF is used for treatment of couples suffering from male factor alone closely followed by cases involving only ovulatory dysfunction alone. Great care should be taken in comparing a reported success rate for type of infertility with individual couple’s own diagnosis. There are very marked variations within each diagnostic category and no two couples are truly the same,


 3.  Age 
Age is probably the single best predictor of IVF success. Individuals over age 42 have markedly reduced chances of success in all clinics. While there is a gradual decline of fertility with age, the greatest shift occurs around age 38. In 2001 the live birth rate fell from 15.9% at age 40 to 5.9% at age 43 and 2.9% at ages above 43. At age 42, the miscarriage rate was 40%, increasing to near 60% at age 44 and over. As the CDC points out, clinics should not be compared by their success rates for women over age 42. There is much greater chance of cycle cancellation and miscarriage once a pregnancy is established in women over age 40.

Clinics vary dramatically in the percentage of their total IVF enrollment above age 38. Clinics also vary in the stringency of their enrollment criteria. It is unclear whether these two facts are related. Certainly, clinics in which patients are highly screened will have better success rates than those clinics with an open enrollment allowing most patients to attempt IVF regardless of anticipated success.

A common diagnostic test and predictor of ovarian responsiveness is the measurement of follicle stimulating hormone (FSH) and estradiol (E2) on cycle day 2-3. Higher FSH levels do not preclude a pregnancy, but they usually predict a decrease in the number of eggs available from ovarian stimulation and therefore, may also indicate decreased success. Clinics vary in the level of FSH considered exclusionary to proceed with treatment.


4. Number of cycles

A cycle is counted when a woman begins taking drugs for ovarian stimulation. In 2001, clinics performing fewer than 60 cycles had a 23.4% chance of pregnancy compared with 27.5% in clinics performing over 240 cycles per year. Clinics performing between 60-248 cycles had a 26.3 % success rate. As a group, larger clinics may have more uniform and predictable, but not better, success rates. Care should be taken when choosing smaller clinics. Some smaller clinics have excellent success rates while rates with others are very poor.

Over 50% of IVF pregnancies are achieved in the first cycle. Another 22% are achieved after one additional try and 13% will occur with three or more attempts. Some clinics may offer multiple repeat cycles to individuals with poorer chances of success, while other clinics are more likely to suggest egg donation.


5. Percentage of cycles resulting in pregnancies / live births 
The difference between the cycles resulting in pregnancies and those resulting in live births represents the miscarriage rate. Approximately 17% of all patients regardless of age or number of pregnancies resulted in an adverse outcome (miscarriage, induced abortion, or stillbirth.) The chances of miscarriage in otherwise fertile women have been variously reported to be between 8-15 %. Pregnancy loss is probably related much more to individual patient characteristics than to the IVF procedure, or to the individual clinic.

While the live birth rate is important in the personal decision to pursue therapy, the IVF program itself is probably better judged by its pregnancy rate. In the past, pregnancy has been defined variously as either a positive hCG level (pregnancy test), ultrasound evidence of a pregnancy or evidence of fetal heart activity on ultrasound. In 2001, pregnancy was defined as ultrasound evidence of a gestational sac. Miscarriage rate probably should not impact selection of an IVF clinic.


6. Confidence Interval
Statistics are a method of evaluating how well a reported value can be trusted to be   correct.

The confidence interval is stated as a range from the lowest possible to the highest possible value that may occur by chance alone. For example, if a small clinic is performing 20 cycles and reports a 40% pregnancy rate, a confidence interval, or true success rate, will range between 20% and 70%, a big difference. A clinic performing 500 cycles may report a 40% pregnancy rate but the range of possibility would be much narrower, possibly 38%42%. The bottom line is that differences of several percentage points between clinics may not prove one clinic superior to another. The difference may not be statistically significant, or occur by chance alone.


7. Percentage of retrievals/transfers resulting in live births 
Fewer than 5% of retrievals do not result in transfers. This may be a result of a variety of factors including the failure to obtain eggs, failure of the eggs to fertilize, or poor development of the resulting embryos. The leading cause of obtaining no eggs at aspiration is failure to properly use the hCG injection. Failure of fertilization can be related to egg quality, sperm quality, or a combination of factors. Laboratory problems have become increasingly rare as a cause of fertilization failure. Certainly, a couple evaluating the decision to proceed to IVF should consider take home baby rate (cycle start to live birth). However, in evaluating an IVF clinic, the most useful parameter is transfer rate to pregnancy rate and this statistic is not specifically given in the CDC data. The reason for this is that individual clinics vary in IVF patient selection criteria and individual patients vary in their chance of miscarriage.


8. Percentage of transfers resulting in singleton live births

See multiple pregnancies below.


9. Percentage of cancellations 
Clinics vary dramatically in their cancellation rate. Cancellation may occur for a variety of different reasons, but the vast majority in 2001 (84.4 %) was for a poor response to ovarian stimulation. About 3% of cycles were cancelled because of over stimulation. Some clinics may require one or more cycles of gonadotropin therapy before IVF.

While these stimulation cycles may have therapeutic value, they also provide a screening test that helps predict the amount of gonadotropin therapy needed in the subsequent IVF attempt, thus reducing cancellation rate. Clinics may also discourage those patients with a less good response with gonadotropin therapy from proceeding to IVF. Some clinics require five or more preovulatory size follicles before proceeding with follicle aspiration while other clinics require only two. Other clinics may move directly to IVF without prior knowledge of gonadotropin responsiveness and convert the cycle to an intrauterine insemination (IUI), which often has a reasonably good chance of success in itself, but these pregnancies are not counted as success in the CDC data. Couples probably should be more concerned with clinics having a relatively low cancellation rate because these clinics may be excluding patients before enrollment in IVF by using overly aggressive stimulation regimens.Overall, cancellation rate is not a very useful predictor of clinic success and may be more a matter of style than substance


10. Average number of embryos transferred/multiple gestation 
In 2001, 27.3% of all transfers involved 2 embryos; 34.5% were of 3 embryos;

20.6% were of 4 embryos, and 11 % were of 5 or more.

The number of embryos transferred represents the second most important clinic statistic after success per embryo transfer. For example:

If clinic A transfers 3 embryos with a pregnancy rate of 40% and clinic B transfers 2 embryos with a pregnancy rate of 30%, it is quite possible that clinic B represents a superior clinic. While pregnancy rate obviously is of great importance, the biggest pitfall of CDC reporting has been the focus on success rate as the ultimate objective.


In 2001, an entry was first made specifically indicating the percentage of singleton births. There is an additional shortcoming of the CDC data in that neither the number of patients delivering triplets or greater nor the number of selective abortions are recorded.


There can be no debate that the ultimate objective should be the birth of a single healthy child. It is also undeniable that even twins significantly increase the risk in a pregnancy. Whether the risk of twin pregnancy is an acceptable risk for the IVF patient has not been established. In our present environment and in keeping with the 2004 ASRM guidelines, there seems little reason to transfer over two embryos in women under age 35 and possibly in all patients. The number of embryos transferred should not exceed the number of pregnancies desired. (Also see the comments below on multiple gestations.)


11.  Cryopreservation 
In general, cryopreservation success has significantly improved over the last ten years. Embryos that would establish a pregnancy probably both freeze and thaw well. Birth rates after cryoprescrvation were 23% per transfer compared to 33.4% in "fresh" cycles. There is no CDC reporting of whether the freeze/thaw transfer was after successful or unsuccessful IVF cycles.


The upside of cryopreservation is that the cumulative pregnancy rate per patient after a single IVF attempt is greater with the costs less than in a single cycle. If fewer embryos are transferred in a fresh cycles there may be more good quality frozen embryos that remain.


Good quality embryos remaining for cryopreservation also indicate a higher chance of pregnancy in a fresh cycle. The down side is that surplus embryos for freezing may indicate an aggressive stimulation cycle and many couples want to limit the family or have ethical concerns about embryos in storage. Frozen embryos may allow for a second pregnancy, and/or and an additional transfer(s). It is uncommon for a single IVF cycle to result in over two total pregnancies, but they both could be multiple. Because of the relatively good cryopreservation rates, there should be less emphasis on number of fresh embryos transferred. There is no substantiation that embryos loose viability the longer they are kept in storage.




What's wrong with the CDC reporting?

Clinic versus individual success rates



The CDC has performed a service in promoting uniformity in success rate reporting. It is ironic that on every page it warns consumers NOT to compare clinics but in effect, comparison is clearly what occurs with the data. Couples should ensure that the success rates quoted by a clinic conform to those uniform-reporting procedures. IVF remains a competitive business with high stakes. Every clinic likes to portray itself, positively but falsifying data is paramount to consumer fraud. The clinic specific reports can help frame one's own decision-making process in terms of choosing IVF as a therapy. Certainly, the clinic should approximate the National success rate average, but in the final analysis it matters less what clinic specific success rale is than that of an individual couple predicted success within that clinic. Frank and open dialogue between patients and their clinic physician is imperative.



Promotion of multiple gestations

 While there are obvious benefits from open competition among centers, the push for high pregnancy rates has placed patients and their pregnancies at significantly higher risk.


The CDC reports have potentially increased competitiveness by their emphasis on pregnancy rate. Clinics may use larger amounts of fertility drugs to avoid cycle cancellation and increase the number of eggs and embryos. Pregnancy rate is directly proportional to the number of good quality embryos transferred. Pregnancy risk is proportionate to the number of gestations. Outside the U.S, the number of embryos that can be transferred is often regulated by law. The American Society for Reproductive Medicine (ASRM) has published voluntary guidelines suggesting that no more than 2 embryos are transferred in women under age 35. A glance at the CDC statistics quickly reveals that few centers are in compliance. Whenever IVF pregnancy rate is reported, the average numbers of embryos transferred must also be given. It seems unconscionable for centers with high pregnancy rates also to have a high average number of embryos transferred.


The high cost, lack of insurance reimbursement, and often self-pay nature of IVF therapy has been used to justify aggressive ovarian stimulation and transfer of multiple embryos to insure the greatest chance of success. Multiple pregnancy risks have been of secondary importance. Several studies have reported that some couples desire multiple births, even triplets or more. Some use selective abortion as a "safety net" to reduce higher order pregnancies to one or two gestations. In reality the financial and/or emotional cost of multiple gestations is substantial. It has been argued that the financial cost for care of multiple gestations paid by insurance companies exceeds the amount it would cost them if IVF were a completely covered expense and transfer number was limited. Insurance companies take note.


Implantation rate

Among embryologists, the most important marker of a successful program is the implantation rate, i.e., the number of embryos divided by the pregnancy rate. Presently this is in the range of 20%. By reporting this figure, it de-emphasizes boosting pregnancy rate by increasing the number of embryos transferred. Implantation rate is a reasonable question to ask of an IVF clinic.


Because the data on pregnancy outcome for any given year can only be completed after the last baby is delivered from the final IVF cycle in that year, there is a long delay between the compiling and reporting of the information. IVF clinics are not aware of statistics on how their clinic compares with the national averages for as much as 3 years.

This may hamper their policymaking.


As this editorial is written it is toward the end of 2004 and I am using CDC statistics from 2001. Undoubtedly the national ART composite has changed. Clearly many programs have changed. There can be major program staff changes, laboratory improvement or deterioration, changes in aggressiveness of stimulation, or number of embryos transferred - all of which can affect success rate. Each clinic may have a more current report available for patients, but then in some ways we are back where we started before mandated reporting.



At its inception, the law mandating clinic specific reporting was for consumer protection. The financial cost of IVF has always been high and an absolute barrier for many. If the CDC report is to be truly consumer oriented, there should also be a listing of the average cost per cycle. It would seem with two clinics (A and B) both having the same services and success rates but A charging 25% more than B --- is not B the better clinic?

Dr. Thatcher was a much loved and longtime friend and supporter of INCIID and Consumers dealing with infertility. He was also an Advisory Board Member of INCIID. His voice continues to be greatly missed.