What does life-table analysis of sequential euploid blastocyst transfers tell us about recurrent implantation failure–or lack thereof? Authors: Richard J Paulson Published: 2026 Life-table analysis is a useful statistical method that allows the cumulative probability of a particular event to be determined. It is used in cancer treatment to estimate the cumulative survival rates associated with a particular treatment, and in fertility treatment to estimate the cumulative probability of success after a number of attempts with a particular therapeutic intervention. Two different treatments may also be compared using life-table analysis. In this manner, small differences that may not be apparent after only one or two treatment cycles become amplified through repetition, so that a larger—and statistically significant—difference is eventually appreciated. The field of infertility medicine embraced life-table analysis in the early days of ovulation induction, using it to estimate the cumulative success rates of several cycles of ovulation induction, intrauterine insemination, or even observation alone among couples with unexplained infertility attempting natural conception. In this way, life-table analyses can provide prognostic estimates for specific treatments in specific clinical situations, such as the monthly probability of pregnancy after two or three years of unexplained infertility. The gradual decline in pregnancy rates over successive cycles can also confirm the worsening prognosis of those who remain in the population. Some 30 years ago, our group performed a life-table analysis of cumulative live birth rates after egg donation, which was then the most successful form of fertility treatment [1]. The analysis demonstrated a live birth rate >80% after five successive egg-donation cycles. Because no decline was observed over five successive treatments, we postulated that “the lack of decline in success during consecutive cycles does not support the concept of the unmasking of an intrinsic endometrial defect in any of the groups considered here. The observed cumulative delivery rate of 86.1% after four cycles of oocyte donation, thus, seems applicable to women of all ages and diagnostic categories.” However, this observation was quite new, and the search for an intrinsic endometrial defect went on. Over the last few years, several reports have assessed cumulative live birth rates after sequential single-embryo transfers, using what is now accepted as the most successful form of fertility treatment: the transfer of preimplantation genetic testing for aneuploidy (PGT-A)–tested euploid blastocysts [2, 3, 4]. Whereas per cycle live birth rates in the 1990s were <30%, (even with the transfer of multiple embryos in a single attempt,) current live birth rates after PGT-A testing exceed 50%. Therefore, it is no wonder that cumulative live birth rates after just three cycles of euploid blastocyst transfer surpass 90% [2, 3, 4]. The per cycle and cumulative live birth (LB) rates for the referenced studies are summarized in Table 1 and depicted graphically in the first line of Figure 1. All of the quoted studies also made the argument that their data implied that an intrinsic block to pregnancy, now called, “recurrent implantation failure (RIF)” was rare. Table 1Summary of the three large studies of euploid embryo transfers conducted to date.Attempt #Pirtea et al. (3)Gill et al. (4)Durbin et al. (1)Total per cycle LBCumulative LB12870/4429 = 64.8%65%3664/5811=63.1%6534/10,240 = 63.8%63.8%2481/885=54.4%54%757/1403=54.0%1238/2288=54.1%83.3%392/170=54.1%44%208/392=53.1%300/562=53.3%92.2%442/105=40%59/117=50.4%101/222=45.5%95.7%524/45=53.3%19/33=57.6%42/78=53.8%98.0%Percentages represent live birth (LB) rates in individual attempts; the cumulative LB rate is shown in the last column. Summary of the three large studies of euploid embryo transfers conducted to date. Percentages represent live birth (LB) rates in individual attempts; the cumulative LB rate is shown in the last column. Life-table analysis of cumulative live birth rates in the summary of actual data in Table 1 and in the four hypothetical situations in Tables 2A–D. Despite these arguments and the high reported cumulative pregnancy rates, there appears to be a continued belief among fertility clinicians that RIF is common, and that it requires investigation and treatment. This false belief leads to many unnecessary and unproven tests for putative endometrial factors that might impede embryo implantation. At a recent medical meeting that I attended, these high cumulative LB rates after euploid embryo transfer were presented, along with the argument that they cast serious doubt on the prevalence of RIF (2, 3, 4, 1). Despite these arguments, many participants were anxious to learn how best to treat patients with presumed RIF. The discussion focused on antibiotics and vaginal probiotics for hypothesized endometritis, despite a lack of evidence that endometritis causes RIF or that antibiotic or probiotic treatment is effective in its treatment. Because the mathematics of life-table analysis is not obvious, it may be helpful to demonstrate how a hypothetical prevalence of RIF would affect individual success rates in sequential embryo-transfer attempts, and then compare these to the actual data. For example, we can demonstrate cumulative success in hypothetical situations associated with a 0% to 20% incidence of RIF. In this case, hypothetical RIF could represent an endometrial block to embryo implantation, such as endometritis, or another unknown factor. The portion of the population with RIF would not conceive without additional interventions. The remainder of the population, without RIF, would continue to experience success according to the probability of implantation. Tables 2A–D demonstrate successive hypothetical treatment attempts in which the incidence of RIF is 0%, 20%, 10%, or 5%. The tables are constructed so that each observes a 50% probability of success in the first attempt. Table 2A assumes a 0% incidence of RIF; therefore, each subsequent attempt has the same 50% success rate. After each attempt, the portion of the population that is successful is removed from the population. Therefore, if we begin with 1,000 subjects, 500 will experience LB at the end of the first attempt and will be removed from the population. The second attempt will then be undertaken by the remaining 500 subjects, and so on. It is easy to show that, in this scenario, 97% of the population achieves success after five consecutive cycles. Table 2Calculated live birth rates in a theoretical population of 1,000 subjects with a hypothetical percentage of recurrent implantation failure (RIF). Subjects with RIF do not achieve live birth (LB), whereas those without RIF have the same success rate in each attempt. The success rate is set at 50% for the first attempt to demonstrate how the curves diverge over time.2A (No RIF; success rate for subjects without RIF = 0.500)Attempt #NNo RIFRIFLB (%)Cumulative LB (%)11,0001,0000500 (50)5025005000250 (50)7532502500125 (50)884125125063 (50)9456363031 (50)972B (20% RIF; success rate for subjects without RIF = 0.625)Attempt #NNo RIFRIFLB (%)Cumulative LB (%)11,000800200500 (50)502500300200188 (38)69331211220070 (22)7642424220026 (11)7952161620010 (5)792C (10% RIF; success rate for subjects without RIF = 0.555)Attempt #NNo RIFRIFLB (%)Cumulative LB (%)11,000900100500 (50)502500400100222 (44)72327817810099 (36)8241797910044 (25)8651353510019 (14)882D (5% RIF; success rate for subjects without RIF = 0.526)Attempt #NNo RIFRIFLB (%)Cumulative LB (%)11,00095050500 (50)50250045050237 (47)74326321350112 (43)8541511015053 (35)90598485025 (26)93 Calculated live birth rates in a theoretical population of 1,000 subjects with a hypothetical percentage of recurrent implantation failure (RIF). Subjects with RIF do not achieve live birth (LB), whereas those without RIF have the same success rate in each attempt. The success rate is set at 50% for the first attempt to demonstrate how the curves diverge over time. Table 2B assumes a RIF incidence of 20%, meaning that, among 1,000 participants, 200 have RIF and 800 do not. If the success rate in the first cycle is 50%, the actual success rate among the 800 subjects without RIF is 62.5%, whereas the 200 subjects with RIF do not succeed. In subsequent attempts, the LB rate drops very quickly from cycle to cycle as subjects without RIF succeed and are removed from the population. After just a few cycles, it is clear that there is a problem in a portion of the population that warrants investigation. This exercise can be repeated for hypothetical RIF rates of 10% and 5%, as shown in Tables 2C and 2D. The numbers in these hypothetical tables can be compared with the actual reported values in Table 1. The results can also be compared visually in Figure 1. The actual data from the three studies that assessed cumulative success rates after euploid embryo transfer are depicted in the top line. No decrease in implantation rate is observed. Had there been even a 5% incidence of RIF, implantation rates would have diminished quickly. This is the key message: there is no decline in subsequent implantation rates, and no RIF is observed in these series. Given statistical variability and the possibility of a type II error, it is not possible to conclude that RIF is absent. However, by simply comparing the results in Table 1 with those in Tables 2A–D, as well as the line graphs in Figure 1, it is possible to say that its incidence must be less than 5%. The scientific search for an understanding of endometrial receptivity, the uterine microenvironment, and the effect of the timing of progesterone stimulation on embryo implantation is very worthwhile. However, diagnosing patients who have failed no more than five euploid blastocyst transfers as having RIF is not evidence-based. It is fair to acknowledge that this concept is not easy to communicate to patients in the clinical setting. We are still confronted by the patient who asks, “Doctor, what are we going to do differently?” (5) Table 1 can be used when counseling frustrated patients to estimate their probability of success in the next euploid blastocyst transfer, and to reassure them that endometrial biopsies, antibiotics, and probiotics are not necessary. For example, after four unsuccessful transfers, when many patients (and practitioners) are wondering if implantation is not possible, we have the opportunity to show patients the results in Table 1. On the basis of the best currently available data, it is accurate to state that the probability of success in the next cycle is 53.8%. Declaration of Generative AI and AI-Assisted Technologies in the Writing Process During the preparation of this work the author used ChatGPT to ensure the accuracy of grammar and spelling. After using this service, the author reviewed and edited the content as needed and takes full responsibility for the content of the published article. Declaration of Interests R.J.P. has nothing to disclose. References