IVM media are designed specifically to support immature cumulus-oocyte complexes not denuded oocytes that have failed to respond to hyperstimulation

To the Editor:
We wish to express concern about the design, analysis, and
conclusions reported in the article by Moschini et al. (1). This study
compares the efficiency of an oocyte IVM medium (Medicult) to
a cleavage-stage embryo culture medium (Sage IVF), for the in vitro
maturation (IVM) of denuded oocytes that have failed to respond to
gonadotropins in a standard IVF cycle (rescue IVM). We have
a number of objections to this study.
The investigators have attempted to examine the efficiency of an
oocyte IVM medium that is not designed and is not appropriate for
culture of oocytes lacking cumulus cells (denuded oocytes). Oocyte
IVM media are intended for the culture of intact cumulus-oocyte
complexes (COC). Because COC contain 1,000-5,000 cumulus
cells, but just one oocyte, IVM media are designed principally to
meet the nutritional needs of the somatic cells. Hence, IVM media
are complex media capable of supporting the high metabolic
requirements of the COC (e.g., very high glycolytic activity),
whereas cleavage media are designed for the low metabolic needs
of the early embryo (e.g., use glucose poorly) (2). As such, one
should not expect IVM media to be suitable for culturing denuded
oocytes, therefore it is perplexing that the investigators would
choose to undertake such a comparison.
This study is substantially underpowered, which has led the investigators
to make a classic type II statistical error in their conclusions.
There are only 24-38 oocytes in each arm of this study. Not surprisingly,
with such low numbers, the investigators have been unable
to detect a statistical difference between treatments. A simple
power analysis (a ¼ 5%, b ¼ 10%) on the germinal vesicle to
metaphase II group reveals that 664 oocytes would have been
required to correctly conclude no actual improvement with the
IVM medium. Clearly such numbers are not practical, therefore
the decision to undertake the trial was flawed. Instead, the investigators
come to the inappropriate conclusion that the IVM medium is
not better than cleavage medium, when in fact, with so few oocytes,
they would be unable to detect such a difference if it were to exist.
Finally, the investigators come to the disturbing conclusion that
the greater potential application of IVM is to rescue the currently
unusable immature oocytes retrieved in standard IVF, rather than
IVM in unstimulated cycles. There is now ample evidence in the
literature that immature oocytes collected in standard IVF cycles
are inherently abnormal. In fact they cite many examples of this
evidence, but not perhaps the most compelling. Jones et al. (3) conducted
a comprehensive microarray analysis of rescue IVM human
oocytes versus in vivo-matured oocytes. The rescue IVM oocytes
showed grossly aberrant gene expression (2,766 genes), consistent
with features such as chromosomal misalignments, alarming rates
of oocyte and embryo aneuploidies (4, 5), and very limited oocyte
developmental competence after IVF. These findings should not be
surprising, as these rescue IVM oocytes have failed to respond to
the ovulatory gonadotropin stimulus in standard IVF. Although
rescue IVM oocytes may be of some use as research material, we
believe that it is inappropriate for the investigators to advocate the
use of rescue IVM oocytes for the treatment of infertility.