The following are excerpted from the Report of the Royal College of Obstetricians and Gynaecologists (UK) Working Party on Ultrasound Screening for Fetal Abnormalities, October 1997 on the Safety of ultrasound in pregnancy:

".....The assessment of the safety of diagnostic ultrasound has been approached from two directions:

• by searching for epidemiological evidence from populations who have been exposed
• by studying the interactions of ultrasound with biological materials in laboratory environments and attempting to extrapolate the results to the real clinical case.

The difficulties with the epidemiological approach are in controls, dosimetry and end-points. The laboratory approach suffers predominantly from the problems associated with interpreting the data in a 'real' clinical context .... .

It appears that the reasons for the concern arose not from new epidemiological data but rather from trends in the characteristics of probes, their outputs and clinical use.

In 1993 the Department of Health, through the Chief Medical Officer's Radiology Advisory Committee, produced a statement which encompassed a number of issues, including the following advice:

• There is no unequivocal evidence of damage to humans from diagnostic ultrasound exposure
• Some diagnostic fields do present a biophysical hazard but the risk is dependent upon the individual
• Current concerns are with pulsed Doppler and the use of contrast agents
• Purchasers should comply with the IEC I 1 5 7 standard
• Adequate training is essential - professional bodies were identified as being the appropriate source of standards
• Ultrasound equipment should be maintained to HEI 98 standard
• There is a need for further research

Thus the trend has been towards gradually increasing anxiety but without any firm evidence of real risk to humans. Throughout, epidemiologists have been hampered by the lack of guidance about what effects to search for. Salvesen and Eik-Nes summarised the end-points of recent clinical studies as being in five categories:

• Childhood cancer
• Dyslexia
• Non-right handedness
• Delayed speech development
• Reduced birth weight.

More recent studies are of greater value than older publications because they relate to the influence of modem equipment and scanning techniques which are well known to be associated with power output levels significantly greater than those used previously.

Childhood cancer

The two largest epidemiological studies have been carried out in the UK retrospectively using data about children dying of cancer in the late 1970s and early 1980s. No association was found between ultrasound exposure and risk of childhood cancer. However, these data are up to 25 years old.

Dyslexia

The idea that ultrasound might lead to an increase risk of dyslexia has been discussed in two studies. In the first, Stark et al examined 425 children aged 7-12 who had antenatal exposure to ultrasound and 381 matched children who had not. They looked at 16 outcomes, one of which was dyslexia as measured by a single reading test and concluded that there was a significant correlation (p less than 0.01). The second study was carried out in Norway by Salvesen et al who examined 2161 children in two randomised controlled trials. Their assessments were more extensive than those of Stark and they found no correlation. In view of the well-publicised problems of defining and assessing dyslexia, it would seem reasonable to place more emphasis on the second study because of its more sophisticated assessment protocols.

Non-right handedness

TI-le same Norwegian study did find a correlation between ultrasound exposure and non-right-handedness. 19% of the exposed children were non-right handed as compared with 15% of the controls. Although this result has been reported as significant, the correlation is relatively poor and is now the subject of ongoing research by the same group.

Delayed speech development

This is another case in which there has been an early positive report and a later negative finding. The original study was performed in Canada and the follow-up was from the Norwegian group. There are numerous criticisms of the original report including its small size, the possibility of bias and poorly matched controls. The second study involved greater numbers, improved randomisation and better controls. In this study, no significant difference was found between children who had been ultrasonically screened and those who had not.

Low Birth Weight

The question of whether ultrasound exposure leads to reduced birth weight has probably been given more attention than any other end point. This may be due to the existence of evidence of such an effect in mice or possibly because it is relatively quick and easy to measure. There are 11 studies to-date. Only three out of the I I studies found a statistically significant relationship between ultrasound exposure and birth weight. Of these three, two found a decrease (Moore et al) of 41 g and Newnham et al 25 g and one found an increase (Waldenstrom et al) of 42g. Moore et al concluded that their findings were probably confounded by other factors, while Waldenstrom et al concluded that there may have been a matemal bias. The data from the study by Newnham et al was actually designed to test a different hypothesis and this has led to considerable criticism. Follow up at one year reveals no weight differences between groups Thus the results of the studies cannot be taken as strong evidence of harm and should not be read in isolation.

Summary of epidemiological evidence

Studies have often been guilty of poor experimental design, poor dosimetry or both. There remains no firm epidemiological evidence of hazard although there is clearly considerable scope for further work-. It would seem logical to look to laboratory studies for further information about interaction mechanisms and possible biological endpoints. Since it is impossible to exclude the possibility of harmful effects, it will never be possible to prove absence of harm. However, the publication of positive instances, detached from the denominator of all instances where the association may have been sought but not found, is likely to give a falsely pessimistic impression of risk. For example, if positive associations, not necessarily hypothesised when the study was set up, are reported more often (in all probability much more often) than negative ones, then false alarms are inevitable from time to time. The only way to tackle this problem would be through large studies in which the associations that were to be looked for were declared in advance, and where statistical adjustment was made for the number of associations sought.

Available data in the literature do not indicate any specific risks of physical damage arising from the use of ultrasound. Cavitation can be produced in vitro by ultrasound but there is no evidence that it can cause human damage in vivo. Microstreaming is an uncertain phenomenon in vivo and may, in any case, not be damaging. Thermal effects are perhaps those of greatest concern and precautions must ensure that significant temperature rises are avoided in practice. Particular care should be taken in certain clinical circumstances, for example early pregnancy, and with ultrasound modes such as pulsed Doppler.

The meta-analyses of randomised controlled trials of adverse effects show only that there is a just significant increased tendency to non-right handedness in the offspring of women who had scans; the complexity of the study makes the observation difficult to interpret. Nevertheless continual vigilance is necessary particularly in areas of concern such as the use of pulsed Doppler in the first trimester.

The greatest risks arise from the use of ultrasound by inadequately trained staff, often working in relative isolation and using poor equipment. Ready referral to a tertiary centre for further advice, usually from a multidisciplinary team, is an essential feature of the screening process ....................".

Also useful for reading are the following websites:

The AIUM statements of safety for patients.
Safety considerations from the British Medical Ultrasound Society.
Safety of Diagnostic ultrasound in Fetal scanning.
Ultrasound Scans Linked to Brain Damage in Babies.
Worrying concerns from Birth.com.au.
On the Safety and Usefulness of Prenatal ultrasound.
Ultrasound, more harm than good?.
The accuracy, safety and pitfalls of prenatal ultrasound by Kmom.
Safety of 3D ultrasound from AskQuestions.org.
4-D Ultrasounds Are Risky Entertainment.
AIUM statements of scanning for Entertainment.
The FDA statement on "Fetal Keepsake Videos".

---

Back to Ultrasound page