The lack of natural growth factors in the fluid in which IVF embryos are grown could have lifelong effects on people conceived this way. That is the implication of a study on mice by Australian researchers, who say preliminary studies of human embryos back their claims.

Although the latest review of IVF safety found no differences between 8-year-olds conceived normally and those conceived by IVF or ICSI, the possibility of long-term health effects still cannot be ruled out.

There are certainly differences early on: singleton babies conceived by IVF or ICSI are more likely to have a low birth weight and to die soon after birth (New Scientist, 23Oct04, p 10).

Could these problems have something to do with the process of IVF or ICSI, such as growing the embryos in a dish for two to five days? Growth media have been developed by trial and error, and contain only a few amino acids and other nutrients. They have been kept simple in the hope of avoiding unanticipated effects.

Yet according to Sarah Robertson's team at the University of Adelaide, at least one growth factor is needed. Her team compared the fate of three groups of mouse embryos: embryos conceived naturally and flushed from the mother's body; IVF embryos grown in a normal culture medium; and IVF embryos grown in a medium containing a growth factor called GM-CSF, which a range of mammals produce, including humans.

The placentas of the mouse embryos grown without GM-CSF were smaller and the pups' birth weight lower compared with the embryos conceived naturally. By adulthood, these mice had grown fatter than the other mice, and the males also had smaller brains.

Adding GM-CSF to the culture medium almost completely eliminated these differences, the team will report in the journal Endocrinology.

The differences between all groups were relatively small, cautions IVF expert Alan Handyside [Univ of Leeds, UK]. It is not clear if mouse studies are relevant to humans. Barry Behr, director of the clinical IVF laboratory at Stanford Univ in California, thinks it is highly relevant. "I've always attributed the problems of IVF to the deficiencies of our culture medium."

But Roger Gosden [dir, research, Center for Reproductive Medicine and Infertility, New York] is more cautious. "You have to be careful," he says. "There are substantial differences between species."

Robertson says preliminary studies show that GM-CSF more than doubles the chances of human embryos surviving to the blastocyst stage, from 35 to 75 per cent. She is working with a Danish company called Medicult on products containing GM-CSF.

But if GM-CSF increases survival by keeping abnormal embryos alive, it could increase the risk of health problems, Handyside points out. He thinks existing growth media do need improving. But he does not believe that experimenting with new components can be justified until researchers know exactly what effect these components have. "Without a scientific basis you cannot anticipate the effects," he says. "That is the bind we find ourselves in."

[http://www.newscientist.com/channel/health/mg18524874.000, 19Feb05; From issue 2487 of New Scientist magazine, page 7; N Valko RN, 20Feb05]