Hundreds of viable wanted pregnancies are being lost to British women each year because of unreliable miscarriage tests, new research suggests.
The precise numbers are unclear but scientists believe there could be 400 errors every year with the potential to result in terminations.
This is more than the estimated 300 cot deaths reported each year in the UK.
Experts today called for more research and improved professional guidelines to minimise the chances of such tragedies occurring.
Pain or bleeding, or a previous history of miscarriage, may prompt doctors to carry out an ultrasound investigation early in pregnancy.
Common practice is to measure the size of the gestational sac and the embryo it contains.
If a sac of a certain size appears empty, or an embryo reaches a certain stage of development with no heartbeat, a diagnosis of miscarriage is made.
When there is doubt about the results, physicians are advised to re-measure the gestational sac seven to 10 days later.
If the sac does not grow during that time, it is assumed a miscarriage has occurred. A termination may then follow.
But according to the new findings, from a series of research papers published in the journal Ultrasound in Obstetrics, the margin of error is so high that some perfectly viable pregnancies are being lost.
One study revealed an up to 20% variation in the size of gestational sacs reported by different clinicians measuring the same pregnancies.
Professor Tom Bourne, from Imperial College London, who led a study of more than 1,000 women suspected of miscarrying, said: "Currently there is a risk that some women seeking reassurance with pain or bleeding in early pregnancy may be told they have had a miscarriage and choose to undergo surgical or medical treatment when the pregnancy is in fact healthy."
Criteria used to diagnose miscarriage varies. In the UK, one classification under Royal College of Obstetricians and Gynaecologists guidelines is an empty gestational sac measuring 20 millimetres or more. Alternatively, a foetal size of 6mm or more with no heartbeat may also indicate a miscarriage.
In the US, a smaller gestational sac threshold of 16mm is adopted. According to the research, this could result in even more wrong diagnoses.
Each year an estimated 500,000 women have miscarriages in the UK. Research suggests around 30% of these women undergo early ultrasound scans.
Prof Bourne and colleagues write in their paper: "Applying a cut-off for MSD (mean sac diameter) of 20mm could lead to 400 viable pregnancies potentially being misclassified, compared with approximately 300 'cot deaths' reported in the UK each year.
"These numbers are significant and relate to pregnancies that would be highly likely to reach term."
The researchers suggest diagnosing miscarriage on the basis of an empty sac measuring 25mm, or a foetal size of 7mm or more with no heartbeat.
Speaking at a briefing in London, Prof Bourne said a repeat scan should be undertaken 10 days after the first "in all cases".
However, he added: "It's the responsibility of the profession to put its house in order."
Another study, led by Dr Shakila Thangaratinam, from Queen Mary, University of London, indicated that current guidelines were based on old and unreliable data.
"The majority of ultrasound standards used for diagnosis of miscarriage are based on limited evidence," she said.
The experts agreed there was a need for new research involving larger numbers of patients.
Dr Tony Falconer, president of the Royal College of Obstetricians and Gynaecologists (RCOG), said: "The RCOG peer reviews its guidelines routinely and works with other organisations such as Nice (National Institute for Health and Clinical Excellence) to develop best practice advice for doctors. The findings from these papers add to our knowledge of clinical practice and will be considered when we update our guidelines.
"Miscarriage is an upsetting experience for anyone and all women who have had a miscarriage should have access to support from the NHS. Healthcare professionals must receive the best training possible to ensure that they are competent in antenatal screening and diagnoses so that mistakes are avoided."
Professor Jane Norman, director of the Tommy's Centre for Maternal and Foetal Health, University of Edinburgh, said: "In women in early pregnancy, it can often be difficult to confidently distinguish a very early ongoing pregnancy from a miscarriage. Most UK units will adhere to guidelines from the Royal College of Obstetricians (RCOG) on the management of early pregnancy loss in order to make this diagnosis.
"These new data show that following the current RCOG guidelines will give the correct diagnosis in the vast majority of women. They also provide some robust new evidence on which to further improve guidelines for diagnosis in this very important clinical condition."
Prof Bourne sought to reassure women who might be worried about the study findings.
"I don't think women should be anxious, but I do think we should get it right so we don't make any mistakes," he said. "What I'm not saying is that it's likely that someone who has had surgery for miscarriage has had a misdiagnosis of miscarriage."