Friday October 14 2011
The studies looked at ultrasound in early pregnancy
Several news sources have today reported that errors during early-pregnancy ultrasounds are leading to unnecessary abortions. The Daily Mail said that hundreds of babies a year may die due to 'blunders' in testing and the Metro said that unreliable tests caused a baby to die every day.
These alarming claims are based on four studies on the use of ultrasound in early pregnancy. The research looked at the fine measurements that are applied to embryos in the first few weeks of pregnancy to determine whether a pregnancy is viable or potential miscarriage.
By examining scans of over 1,000 women the researchers found that under current guideline measurements around 0.4-0.5% of pregnancies that develop normally would have been misdiagnosed as non-viable. However, by slightly increasing the maximum sizes used to indicate a viable pregnancy, the researchers found that there were no cases of a viable pregnancy being misdiagnosed as a miscarriage. They also say that a repeat scan should take place if there is any doubt.
Despite what headlines have suggested, early pregnancy ultrasounds are invaluable and highly accurate diagnostic tools, and even using current guidelines, the vast majority of cases would be accurately diagnosed. However, this study has demonstrated there is room to further improve guidelines and the care of women in early pregnancy.
It should be noted that while the researchers did suggest there may be around 400 cases of misdiagnosis each year, there is no indication that the majority of them would be terminated, as newspapers have reported.
Where did the story come from?
The news is based on four studies examining the use of ultrasound scanning to monitor early-stage pregnancies. The studies were performed by researchers and doctors from a number of institutions, including Imperial College London, hospitals within the Imperial College NHS Trust, Queen Charlottes and Chelsea Hospital and the KU Leuven University in Belgium. The research was funded by Imperial College and the NIHR Biomedical Research Centre. The studies were published simultaneously in the peer-reviewed medical journal Ultrasound in Obstetrics and Gynecology.
Press coverage of these studies tended to be quite alarming, with suggestions that 400 babies a year die or are terminated due to errors in testing. For example, the Metro ran front-page coverage saying that ‘a baby per day dies due to test error’, while the Daily Mail said that fears were ‘hundreds of healthy babies are being aborted every year simply because of scan blunders’.
However, this figure appears to be based on one research paper that estimated around 400 UK pregnancies may be misclassified as miscarriages, which does not necessarilly mean they are terminated.
At a press conference attended by reporters from a number of national newspapers, some of the study authors stated that there was no reliable source of evidence to confirm how often misclassified pregnancies would be terminated. However, they did say that there had generally been a move towards a ‘wait and watch’ approach, with doctors tending to wait and confirm the diagnosis rather than performing surgery or a termination.
Also, many press stories were accompanied by pictures of late-stage ultrasound scans, showing clearly visible foetuses. This suggests that doctors are performing terminations close to the natural end of pregnancy, when in fact these studies were concerned with diagnosis within the early stages of pregnancy, when an embryo might typically be around 5-6mm in length.
What kind of research was this?
These were four related studies examining the use of ultrasound to diagnose miscarriage. However, due to the length and complexity of these studies, this Behind the Headlines article mainly discusses the particular study called Limitations of current definitions of miscarriage using mean gestational sac diameter and crown-rump length measurements: a multicenter observational study. This was a cross-sectional observational study that collected data on women who were scanned in early pregnancy at four London hospitals.
Women will be offered an early ultrasound if they experience lower abdominal pain, vaginal bleeding, have a poor obstetric history, or to estimate the gestational age of their baby. Early ultrasounds assess particular symptoms or situations, and differ from the standard antenatal screening ultrasound typically given at 10-14 weeks of pregnancy.
To estimate whether a miscarriage has occurred, the health professional performing the ultrasound will look at a number of measurements, including the average length of the ‘gestational sac’ an embryo will grow in and the length of the embryo from the crown to the rump.
To estimate whether a miscarriage has occurred, the health professional performing the ultrasound will look also at the size of the gestational sac when no embryo can be seen, and if an embryo can be detected, its length from the crown to the rump if no heartbeat can be detected.
Current guidelines issued by the Royal College of Obstetricians and Gynaecologists state that a miscarriage may be diagnosed if an ultrasound scan inside the vagina identifies an empty gestational sac with a mean diameter of 20mm or more, or an embryo with no detectable heartbeat with a crown-rump length of 6mm or more. An empty gestational sac of less than 20mm is defined as an intrauterine pregnancy of uncertain viability, and a repeat scan at a minimum interval of one week is advised, although the criteria used to define miscarriage at the repeat scan is not defined. This guidance is based on expert opinion.
There is considerable variability in the criteria used to diagnose miscarriage worldwide, and a number of studies have proposed different cut-off values. For example, in the US an empty sac with a diameter of just 16mm is considered to indicate miscarriage. The researchers aimed to establish cut-off values that can be used confidently to classify a non-viable pregnancy.
What did the research involve?
The study enrolled 1,060 consecutive women who had had an early scan and had been diagnosed with a pregnancy of uncertain viability. A pregnancy of uncertain viability was defined as:
- an empty gestational sac with or without a yolk sac but no embryo and a mean gestational sac diameter of less than 20mm or 30mm (depending on hospital), or
- an embryo with no heartbeat and a crown-rump length of less than 6 or 8mm (depending on hospital)
These women then had another ultrasound 7-14 days later, and another at the time of normal first-trimester screening (between 11 and 14 weeks), to determine whether they had a viable pregnancy or not. The researchers used this data to determine the effects of different cut-off values.
The researchers also recorded other variables if they had had infertility treatment, including the date of last menstrual period or known date of conception. They recorded symptoms such as vaginal bleeding with or without clots and pain.
They followed up women to find out the viability of the pregnancy at 11-14 weeks, which was the main outcome of the study. It was recorded at the time the women were given their routine nuchal translucency scan, a screening test for Down’s syndrome.
What were the basic results?
Of the 1,060 women with a pregnancy of uncertain viability at the early scan, 473 (44.6%) went on to have a viable pregnancy and 587 (55.4%) were found to have a non-viable pregnancy on the later scan. The researchers then examined the accuracy of various different diagnostic criteria:
- If a cut-off value for mean gestational sac diameter of 20mm had been applied to pregnancies where the yolk sac and embryo had not been visualised, 0.5% of pregnancies would have been incorrectly diagnosed as non-viable (one pregnancy in the context of this study).
- Similarly, when a cut-off value for mean gestational sac diameter of 20mm had been applied to pregnancies where the yolk sac had been visualised but the embryo had not, 0.4% of pregnancies would have been incorrectly diagnosed as non-viable (one pregnancy).
- Either with or without yolk sac there were no cases of a viable pregnancy being misdiagnosed as a miscarriage when a cut-off value for mean gestational sac diameter of 21mm or more was applied.
- When an embryo was visible with absent heartbeat, there were no viable pregnancies misdiagnosed as miscarriages when a cut-off for crown-rump length of 5.3mm or more was applied.
How did the researchers interpret the results?
The researchers extrapolate their results using unpublished data of a recent survey conducted by the association of early pregnancy units in the UK. This survey suggested that 500,000 women attend these units each year, and 16% of them have an empty gestational sac of less than 20mm. The researchers say that using their rates of misdiagnosis of viable pregnancies as miscarriages, applying a cut-off of 20mm could lead to 400 viable pregnancies being misclassified as miscarriages.
However, in a press conference on the matter, some of the studies’ authors have discussed the difficulties in accurately estimating the numbers involved. This is due to there being no central register routinely recording data on the matter and because women with symptoms warranting an early scan might present themselves to other medical units, such as accident and emergency or their GP.
The researchers also drew upon data from another study from this series. This study found that there were differences of ±18.78% in measurements made by two skilled examiners, meaning that a measurement of 20mm by one examiner could be read as a measurement of between 16.8mm and 24.5mm by a second examiner. While newspapers have suggested this variation was due to a lack of care or skill, (referring to them as ‘blunders’ by people performing the ultrasounds) the research paper did not suggest this variation was due to a lack of diligence. Instead it looked at the implication of variance that can occur between skilled practitioners.
The researchers have suggested a new set of ‘safe’ cut-off values. They did this drawing upon this study and the findings that there were no viable pregnancies misdiagnosed as miscarriages when using cut-off values for mean gestational sac diameter of 21mm or more without an embryo and crown-rump length of 5.3mm or more of embryos without a detectable heartbeat. These recommend miscarriage be diagnosed when there is a mean gestational sac diameter of 25mm with no embryo visible or when an embryo without a detectable heartbeat has a crown-rump length of 7.0mm or more. They also say that repeat scans should be performed if measurements are close to the cut-off values.
The researchers also say that there should be further clarity regarding what to expect on repeat scans. They refer to another study they have published on this issue, which found that it is possible a viable pregnancy could show no growth in mean gestational sac diameter over 10 days, and that there were no viable pregnancies when a repeat scan found that the gestational sac was still empty with no yolk sac or embryo present.
The researchers report that women who had been scanned because of suspected miscarriage can be managed without the need for medical treatment and surgery. They say that, ‘waiting 7-10 days in order to repeat a scan is highly unlikely to lead to physical harm. The anxiety associated with being uncertain about the status of a pregnancy is very significant, but should be balanced against the possibility of inadvertent termination which is surely the worst possible outcome for any women.'
This research looked at the processes used to diagnose whether a pregnancy was viable when performing an early pregnancy scan. These types of scans were given to women who had an early pregnancy scan due to lower abdominal pain, vaginal bleeding, poor obstetric history or to estimate gestational age. A diagnosis of uncertain viability was made if no embryo could be seen although the gestational sac was a certain diameter, or if no foetal heartbeat could be detected in embryos of a certain size. A subsequent first-trimester scan was performed at 11-14 weeks. The researchers then analysed the results to determine how many viable pregnancies would have been misdiagnosed as miscarriages using the current recommended cut-offs.
Using the current cut-off, the researchers found that between 0.4-0.5% of viable pregnancies would have been misdiagnosed using a cut-off for mean gestational sac diameter of 20mm or more without an embryo, which is the value often used in clinical practice. However, they found there were no cases of a viable pregnancy being misdiagnosed as a miscarriage when they applied a revised cut-off value of 21mm or above. Also, no viable pregnancies would have been misdiagnosed using the current cut-off of 6mm or above for the crown-rump length of embryos with no heartbeat.
However, the researchers also take into account the findings of another study that discovered that measurements between two skilled examiners could differ by ±18.78%. Given this variance, the researchers propose the use of increased cut-off values of 25mm for mean gestational sac diameter without embryo of 25mm and a crown-rump length of 7.0mm or more when no heartbeat is detected to prevent any viable pregnancies being misdiagnosed. They also say that, ‘waiting 7-10 days in order to repeat a scan is highly unlikely to lead to physical harm’.
It should be emphasised that the correct diagnosis would have been made using current guidelines in the vast majority of cases. Professor Siobhan Quenby, talking to the BBC, also said that only 30% of women with a miscarriage diagnosis would take tablets or have surgery to end the pregnancy. The current study and the other studies published in this issue of Ultrasound in Obstetrics and Gynecology will help the current guidelines to be refined. The National Institute for Health and Clinical Excellence (NICE) is currently reviewing its guidelines for pain and bleeding in early pregnancy (due November 2012).