Embryos of good and poor prognosis or of younger and older women are known to behave differently during in vitro culture. The aim of this study was, therefore, twofold: (i) to assess whether a TLI system achieves similar IVF outcomes to standard manual embryology; and (ii) whether the efficacy of a TLI system is the same in better and poorer prognosis patients.
Our center does not change laboratory pratices without prior assessments of non-inferiority and, hopefully, determination of superiority. Before purchasing a closed incubation system, we, therefore, contacted different manufacturers in attempts to perform a prospectively randomized pilot study to assess how such a system would perform in our center’s highly adversely selected patient population. Only one, the manufacturer of the EmbryoScope™, graciously agreed to provide us with a loaner instrument for a 3-months study period. Intarmural research funds were used to pay for installation of the instrument, staff training and supply costs. The study only commenced, once the manufacturer was confident that our center’s embryology staff was competent in using the instrument.
Limited time availability of the system resticted the number of patients we were able to investigate. Here presented conclusions, therefore, should be considered as preliminary, even though they in important aspects are based on statistically significant results.
Prospective randomization of poor prognosis patients (Part A)
Part A prospectively investigated randomized patients with relatively poor prognosis between tradional embryology and the EmbryoScope™. When our investigation was initiated, the instrument had been investigated in only one RCT, involving very favorably selected patients That study reported the instrument to marginally improve IVF outcomes (27). As effectiveness of some IVF interventions differs between good-, intermediate- and poor-prognosis patients [
31], we felt the need to assess its efficacy in our patient population before committing to a purchase. Ages, FSH and AMH levels of here investigated patients (Table
1) reflect the poor outcome prognosis of our patients in comparison to the Spanish study by Rubio et al. [
27].
That in contrast to their study no outcome benefits from the EmbryScope™ were seen in our investigation is, therefore, noteworthy. Our findings may suggest that closed incubation systems in different patient populations may demonstrate different degrees of efficacy. The small size and, therefore, inadequate power of our study, however, does not preclude a Type 2 error in failing to demonstrate an outcome difference in Part A of this study. Significant outcome differences may only become apparent with larger patient numbers. Since observed implantation and pregnancy rates actually trended toward traditional embryology (Table
3), a potential outcome benefit from the instrument over standard embryology, however, appears unlikely.
No differences (ie, no improvements) in embryo development with TLI systems have been reported before [
19‐
21,
28‐
32]. One study, indeed, suggested increased miscarriage rates with use of a TLI system [
26]. As also suggested by Racowsky et al. [
11], whether TLI systems really offer outcome benefits over standard embryology, therefore, remains questionable,
Randomization of embryos in best prognosis patients (Part B)
Part B of this study was intended to assess the instrument in “best” prognosis patients. In using young oocyte donors, large oocyte yields allowed open randomization of embryos between EmbryoScope™ and standard embryology in place of patient randomization. In the previously noted Spanish RCT, oocyte donors represented almost half of the patient population [
27]. The intent, therefore, was to investigate performance of the TLI system in best-prognosis patients, even exceeding the favorable patient selection of the Spanish study.
Since based on the Spanish study we expected improvements in IVF outcomes, we were surprised to observe the opposite. We were further surprised by the statistical power of observed outcome differences, even though small numbers call for caution in interpretation: Despite the relative small study size (Table
5), the EmbryoScope™ produced significantly fewer Grade A and, therefore, significantly more Grade B embryos than standard embryology, suggesting a potentially negative impact of the TLI system on embryo quality.
Observing these rather surprising results raised the question whether staff members operating the EmbryoScope™ had been sufficiently trained. We, therefore, reviewed embryology staffing records, and found that the embryologists handling the instrument during Parts A and B of here reported study were exactly the same. Insufficient staff training, therefore, only unlikey explains the findings of this study.
Our suspicion then fell on the culture dish of the EmbryoScope™, called the EmbryoSlide™. It is a single-use, sterile culture dish, especially designed for the EmbryoScopeTM incubator. Each EmbryoSlide™ holds up to 12 embryos, each cultured individually in droplets of 25 μl media. They, therefore, are cultured in relative low density and not grouped together. In contrast, conventional dishes, as utilized at our center, culture up to 5 embryos in 50 μl droplets.
Higher embryo density group culture has been reported to benefit embryo development [
16,
33], possibly the consequence of one or more factors produced by embryos, which can stimulate embryo development [
34]. We, therefore, parallel cultured alternating donor embryos in EmbryoSlides™ and standard embryo culture dishes in the same incubation environmentone (Table
6). As the table demonstrates, this additional prospective evaluation by day-3 of culture demonstrated no difference in embryo quality between both embryo culture dishes. The EmbryoSlide™, therefore, apparently was not the cause of our observation.
This left us with no established cause for our observation, and the conclusion that the incubation envrironment in the EmbryoScope™, likely, was inferior to our center’s standard embryo culture environment. Such an explanation is not farfetched: Several physical factors can impact final embryo development during in vitro culture, including incubation volume/embryo density, temperature/pH and light as well as shear stress from mechanical motion [
34]. Even minimal changes in temperature during culture (away from 37 °C) for short time periods can result in unrecoverable damage by hurting the stability of the oocye/embryo spindle [
35,
36]. The pH of medium also plays a crucial role in embryo culture [
34], while regulation of pH mostly relies on the carbon dioxide (CO
2) concentration.
The user manual of the EmbryoScope™, notes that the instrument is equipped with unique temperature controls, characterized by direct heat transfer to individual media-filled wells. Temperature is alleged to be virtually unchanged by opening the chamber (<0.2 °C) when adding or removing embryos. Recovery of CO2 concentrations is alleged to occur in less than 5 min and of O2 in less than 15 min after closing of chamber. During this study, we. indeed, based on the built-in monitoring software, did not note temperature and/or CO2 concentration changes of any significance.
It has been suggested that prolonged light exposure required for time-lapse photography may negatively affect embryos [
14,
20,
37]. The EmbryoScope
TM, however, uses long wave length light of lower intensities (red light, 635 nm) than the light used in our embryology laboratory during standard embryo assessments under a microscope (15 % <550 nm). There also are currently no experimental data in the literature to support any negative effects of greater light exposure on embryos with use of the EmbryoScope
TM [
20,
26].
All of this leaves only one likely explanation for our findings: The EmbryoScope
TM contains a microscope, built into a compact incubator. During image capture, the microscope is fixed, while the tray and culture dish move slightly to focus each single well. Since in our study, images were taken every 10 min, embryos during three days of culture had to move at least 380 times. Though this embryo motion is very gentle and mild, the impact of possible sheer stress cannot be ruled out as a potential cause for observed declines in embryo quality [
34].
The effect may be statistically more apparent in best prognosis patients because even a relative small percentage loss in pregnancy in such a population may be statistically more apparent. This also would explain why the reported improvement in outcomes in the Spanish study [
27], which included only a little less than half of best prognosis patients (ie, egg donors) was only marginal. If our here laid out assumptions are correct, had the Spanish study included even more oocyte donors (ie, good prognosis patients), it too, may no longer have demonstrated marginally better IVF cycle outcomes with the TLI system or might even have drifted into the negative.
Mouse models that investigated the effects of mechanical vibration induced shear stress on embryo development, demonstrating decreased morula and blastocyst formation [
38], caused by phosphorylating mitogen-activated protein kinase (MAPK) 8/9 [
39]. Similar phosphorylation of MAPK 8/9 was never observed in control embryo or in vivo cultured embryos.