The suggestion follows new studies which show that one IVF system for infertile men could impair their children's mental development at an early age, although no clear effect has been observed in older children.
However, British doctors played down the implications of the work, saying that unpublished research on children here shows that the treatments - which can cost pounds 2,000-5,000 per attempt - do not pose any risk to a baby's health.
Since IVF was introduced in 1978 to produce the first "test-tube baby", by bringing sperm and eggs together outside the human body, technologies have become increasingly sophisticated.
One, introduced in 1992, allows a single sperm to be injected directly into the egg - virtually guaranteeing fertilisation. Such "intracytoplasmic sperm injection" (ICSI) methods have been used to produce thousands of children worldwide, according to a study in the Lancet medical journal, published today.
But a team of Australian researchers have found that children, and especially boys, conceived by ICSI are noticeably more likely to be "mildly or significantly" delayed in mental development at one year old than those produced by normal conception.
By contrast, a larger, separate study by Belgian medics found that there was no difference in mental development between ICSI-conceived children and the population at large. Belgium pioneered ICSI treatment, and now 3 per cent of all pregnancies there are a result of a form of IVF, including ICSI.
But Edward te Velde, a Dutch fertility specialist, argues in the Lancet that "ICSI ... [circumvents] the selective barriers developed during evolution". He points out that while it initially only used ejaculated sperm, nowadays immature sperm are also used after being extracted surgically from the testes. Such developments could have subtle genetic effects, especially on the X and Y sex chromosomes where abnormalities are known to be more common among infertile than fertile men.
Dr te Velde says, "The findings so far carry a serious warning" and adds, "Before new assisted-reproduction techniques are adopted as routine treatment for infertility, they should be assessed extensively in animal and human embryo research, then in clinical trials, during which the children must be monitored over the long term."
In response, Bert Stewart, scientific director of Midland Fertility Services, said that early results of children older than one in Britain indicated that ICSI did not affect their development. Peter Bromwich, head of the clinic, added: "The biggest problem with IVF is that it tends to cause multiple pregnancies, which leads to premature birth. Being two months premature can make a big difference in a child's development aged one, but mean a lot less by age two."
The Human Fertilisation and Embryology Authority (HFEA), which licenses reproductive technologies in Britain, said: "The difficulty with these techniques is that you can't do comprehensive follow-up studies until the children grow up, and they won't be born unless you let people use the technique. The HFEA was widely criticised for not allowing ICSI in the UK when you could get it elsewhere in Europe."Reuse content